Memory Types | Computer Architecture

Computer Memory Types – Memory is an essential component of any digital system since it is needed for storing programs and data. Most of the general-purpose systems would work more efficiently if they were equipped with additional storage capacity beyond the primary memory. There are several types of memories that are attached to the digital systems.

Types of Computer Memory

1. Magnetic Disk and Magnetic Tape
2. Magnetic Core Memory
3. Binary Cell
4. Read Only Memory (ROM)
5. Flash Memory
6. Random Access Memory (RAM)

Magnetic Disk and Magnetic Tape

A magnetic disk is a circular plate constructed of metal or plastic coated with magnetized material. A magnetic disk uses both sides to store the data. Several disks may be arranged on one spindle with read/write heads available on each surface. All disks rotate together at high speed and are not stopped or started to access the information.

The data is stored in a magnetized surface in sectors along concentric circles called tracks. The data or information is stored in terms of bits. The minimum quantity of information that can be transferred is a sector. The subdivision of one disk surface into tracks and sectors is shown in the figure below.

Magnetic tape is a medium for magnetic recording, generally consisting of a thin magnetizable coating on a long and narrow strip of plastic. Nearly all recording tape is of this type, whether used for recording audio or video or for data storage. Magnetic tape allows a massive amount of data to be stored in digital systems for a long period of time and rapidly accessed when needed. Bits are recorded as magnetic spots on the tape along several tracks. Usually, seven or nine bits are recorded simultaneously to form a character together with a parity bit. Read/write heads are mounted in each track so that the data can be recorded and read as a sequence of characters. The magnetic tape recording mechanism is shown in the figure below. 

Magnetic Core Memory

The magnetic core memory is also known as ferrite-core memory. It is an early form of random-access computer memory. It uses small magnetic ceramic memory. It uses small magnetic ceramic rings, the cores, through which wires are threaded to store information via the polarity of the magnetic field they contain. Such memory is often just called core memory, or normally core.

The most common form of magnetic core memory is X/Y line coincident, current memory. It is used for the main memory of a computer. It consists of a large number of small ferrite rings, named as core. These cores are held together in a grid structure with wires woven through the holes in the core’s middle. In early systems, there were four wires: X, Y, sense, and inhibit. The latter cores combined the latter two wires into one sense/inhibit line. Each ring stores one bit, i.e., 0 or 1. One bit in each plane could be accessed in one cycle. The manipulation of bits on a parallel plane allows the full word to be read or written in one cycle 

Wires pass through the cores create the magnetic fields. Only a magnetic field greater than a certain intensity can cause the core to change its magnetic polarity. To select a memory location, one of the X lines and one of the Y lines are driven, and half the current is required to cause this change. Only the combined magnetic field generated where the X and Y lines cross is sufficient to change the stage; other cores will have only half the field needed. A picture of core memory is shown in the figure below.

Binary Cell

An elementary unit of a digital system that can have one or the other of two stable states and can thus store one bit of information is named a binary cell. The binary cell is capable of storing one bit of information. In a digital system of flip-flop or a latch can serve as a binary cell.

A latch is a kind of bistable multivibrator that has two stable states and thereby can store one bit of information. The simplest type of latch is the set/reset (SR) latch. It can be constructed from either two NAND gates or two NOR gates. The SR latch is shown in the figure below.

Latches and flip-flops are nearly the same, with the basic difference that input and output events are the same in a latch, whereas in a flip-flop, these are two separate events. The flip-flop is a non-transparent (clocked or edge-triggered) device, while the latch is simply a transparent (without a clock signal) device. A flip-flop is usually controlled by one or two control signals and/or a clock. The basic SR flip-flop is shown in the figure below.

Read-Only Memory (ROM)

In this type of memory, data or information is stored permanently. It is a non-volatile memory, which retains stored data or information even if the power is turned off. This memory is used to store programs and data that need not be altered. As the name suggests, the information can be read only. It means once a bit pattern is stored, it is permanent or at least semi-permanent. The concept underlying the ROM can be explained with the diodes arranged in a matrix format, as shown in the figure below. 

In the ROM cell shown in the figure above, the horizontal lines are connected with vertical lines only through diodes. The presence of a diode stores 1, and its absence stores 0. When a register is selected, the voltage of that line goes high, and the output lines, where diodes are connected, go high. For example, if R3 is selected, the data byte 10101010₂ can be read at the data lines D₇ – D_0. This information is permanent in that register.

Generally, ROM is used to store the binary codes for the sequence of instructions. There are four types of ROM: Masked ROM, PROM, EPROM, and EEPROM.

1. Masked ROM: In masked ROM, the data is physically encoded in the circuit, so it can only be programmed during fabrication. It is an expensive and specialized processed ROM. It is impractical for research and development. 
2. PROM: A programmable read-only memory (PROM) is a form of digital memory where the setting of each bit is locked by a fuse or antifuse. Such PROMs are used to store programs permanently. The basic difference from a strict ROM is that the programming is applied after the device is constructed. This memory is suitable for product development, experimental projects, and college laboratories. 
3. EPROM: An EPROM, or Erasable Programmable Read Only Memory, is a type of memory chip that retains its data when its power supply is switched off. It is an array of floating-gate transistors individually programmed by an electronic device that supplies higher voltages. Once programmed, an EPROM can be erased only by exposing it to strong ultraviolet light. EPROM is easily recognizable by the transparent fused quartz window at the top of the package through which the silicon chip is visible. This window permits exposure to UV light during erasing.
4. EEPROM: The EEPROM (also written as E2 PROM) stands for Electrically Erasable Programmable Read-Only Memory. It is a type of non-volatile memory and is used in systems to store a small amount of data that must be saved when power is turned off. This memory allows selective erasing at the register level rather than erasing all the information. The EEPROM memory has a special chip erase mode by which the entire chip can be erased at a fast speed. This is a semi-permanent type of ROM.

There are two types of electrical interface to EEPROM devices: (a) serial and (b) parallel. In microprocessor-based systems, software update is a common occurrence. If EEPROMs are used in the systems, they can be updated through electrical interfaces.

Flash Memory

Flash memory is a non-volatile memory that can be electrically erased and reprogrammed. It is a technology that is primarily used in memory cards and USB flash drives for general storage and transfer of data between different digital systems. It is a specific type of EEPROM that is erased and programmed in large blocks. In early flash memories, the entire chip had to be erased once.

Flash memory costs less than byte programmable EEPROM and therefore, has become the dominant technology wherever a significant amount of non-volatile, solid-state storage is needed. 

Since the flash memory is non-volatile, no power is needed to maintain the information stored in the chip. In addition, flash memory offers fast access time and better kinetic shock resistance. Another feature of flash memory is that when packaged in a memory card, it is enormously durable, being able to withstand intense pressure, extremes of temperature, and even water immersion. A flash memory cell is shown in the figure below. 

Random Access Memory (RAM)

Unlike ROM, random access memory (RAM) can be used for reading as well as writing data. The RAM allows stored data to be accessed in any order. RAM is a volatile memory. There are two types of RAM: (1) Static RAM (SRAM) and (2) Dynamic RAM (DRAM). 

Static RAM (SRAM)

In static RAM, the storage cells are small transistor circuits. These have been used for high speed CPU since the 1950s. After the advent of VLSI Design technology, SRAM is used in the implementation of main memory. SRAM is more expensive but faster, and significantly less power is required. It is therefore used where either bandwidth or low power, or both, are principal considerations. SRAM is also easier to control. There are two ways to design SRAM using transistor circuits: (i) Bipolar memory cell and (ii) MOS memory cell. 

Bipolar Memory Cell: Bipolar static RAM is a static flip-flop type cell. The figure below shows the basic structure of SRAM using bipolar transistors. The bipolar memory cell has two transistors, T0 and T1. These two transistors behave as cross-coupled NAND gates. At any instance of time, one of the transistors is switched on, that is, conducting current, while the other is switched off. There is one address line and two bit lines (b and b), which are used to select a read or write operation by changing the voltage across them. There are three distinct states, represented by three voltage levels: V₀, V_1/2, and V₁. The voltage level V₀ is low voltage, V₁ is high voltage, and V_1/2 is intermediate voltage. 

In normal conditions, both bit lines (b and b) are kept at the V_1/2 voltage level, and the address line is kept at a slightly higher voltage. This condition puts diodes D₁ and D2 in reverse bias, and the memory cell is isolated from reading and writing operations. This is named as stand by operation of the memory cell.

For the reading operation, the bit lines are kept at the V_1/2 voltage level, and voltage across address line is raised. This results high emitter state of the one transistor and causes a current flow through this transistor. Now the sense amplifier detects the state of cell through bit lines. For writing operation, the data (0 or 1) which is to be written, is applied to the bit lines. To write a 0 in cell apply b = 0 and b = 1. Similarly to write a 1, apply b = 1 and b = 0. The memory cell acts as a latch and when the address line is activated, it stores the data in it.

MOS Memory Cell: The basic limitations of bipolar memory cell are cost and power consumption. These limitations can be reduced by using MOS memory cell. The simplest form of MOS memory cell is shown in below figure.

As shown in the above figure, the MOS memory cell has six transistors. Out of these six, four transistors (T₁, T₂, T₃ and T₄) are used for storage purpose whereas remaining two transistor (T₅ and T₆) are used to control the access of cell.

If the address line of the cell is deactivated, the transistors T₅ and T₆ are turned off. This makes the cell isolated from the bit lines. This condition of memory cell is known as stand by condition.

For reading operation, both the bit lines (b and b) are made to high (‘1’) When the address line is activated, it enables both the access transistors (T₅ and T₆). Now the stored value propagates to the bit Iine b and complimented value to the bit line b. The sense amplifier reads the value from bit line.

The writing operation starts with the applying data value (0 or 1) to the bit line b and complemented form to the bit line b. Then address line is activated. This action stores the data value from bit line to the memory cell.

Dynamic RAM (DRAM): Dynamic RAM stores the data as a charge on the capacitor. DRAM contains more memory cells as compared to SRAM per unit area. DRAM is usually arranged in a square array of one capacitor and one transistor per cell. The access time for DRAM is greater than SRAM. The disadvantage of DRAM is that it needs refreshing of charge on the capacitor after every few milliseconds. The single transistor DRAM cell is shown in below figure.

For writing operation, a voltage corresponding the data value (high for ‘1’ and low for ‘0’) is applied at the bit line. The address line is then activated which turns on the transistor T. This causes a charge to be transferred to capacitor C if the bit line is in the ‘1’ state, no charge is transferred otherwise.

For reading operation the transistor is turned on through address line. This results the transfer of charge stored in capacitor C, if any, to the bit line. Now this charge is detected by the sense amplifier. This reading operation is destructive in nature because after reading capacitor is discharged. This condition makes the requirement of refreshing operation (rewriting) into cell

There are various types of DRAM available Few of them are as follows:

SDRAM (Synchronous DRAM): SDRAM is dynamic RAM that has synchronous interface. Traditionally, DRAM has an asynchronous interface which means that it responds as quickly as possible to change in control inputs. SDRAM has a synchronous interface, meaning that it waits for a clock signal before responding to control input and it therefore synchronized with the computer system bus SDRAM can accept one command and transfer one word of data per clock cycle.

RDRAM: Rambus DRAM: RDRAM is a type of SDRAM. It is designed by Rambus Corporation. The RDRAM is a memory that communicates at very high rate of speed. Conventional memory systems those are designed with SDRAM are known as wide channel system. On the other hand, RDRAM is narrow channel device. This transfers data only 16 bits (2 Bytes) at a time but at much faster speed.

A single rambus memory channel can support up to 32-individual RDRAM chips and more if buffers are used. Each individual chip is serially connected to the next on a package called a Rambus Inline Memory Module (RIMM). Another important feature of RDRAM is that it is a low power memory system.

DDR RAM (Double Data Rate Random Access Memory): DDR RAM is better known as DDR SDRAM (Double Data Rate Synchronous DRAM). It is base on the same architecture as SDRAM but utilizes the clock signal differently to transfer twice the data in the same amount of time.

In a computer system, the clock signal is an oscillating frequency used to coordinate interaction between digital circuits. SDRAM operates on only rising edge of the signal to transfer data, whereas, DDR RAM transfers data on both rising and falling edges of the clock signal. Hence DDR RAM is essentially twice as fast as SDRAM.