We begin this chapter with a survey of semiconductor main memory subsystems, including

ROM, DRAM, and SRAM memories. Then we look at error control techniques used to enhance
memory reliability.

Semiconductor Main Memory

In older computers, the main memory used a technology called magnetic core memory. This
memory was made from small, doughnut-shaped magnetic rings called cores, which could store
information. That’s why people used to call memory just “core,” and sometimes still do today.

However, with the invention of microelectronics (small electronic circuits), a better and faster
way to build memory was found. This new method uses semiconductor chips, which are now
used almost everywhere for main memory in computers.

Memory Organization

The smallest building block of this memory is called a memory cell. Each memory cell can store
just one bit of data, which is either a 1 or a 0. Even though there are different types of memory
cell technologies, they all have some things in common:

●​ Each cell has two stable states (either on or off), which are used to show 1 or 0.​

●​ You can write into the cell, which means setting it to 1 or 0.​

●​ You can also read the value from the cell to know if it is storing a 1 or a 0.

How a Memory Cell Works (Refer to Figure 5.1)

In Figure 5.1, you can see how a memory cell works. Usually, it has three important terminals
(or connection points) that carry electrical signals:

1.​ Select terminal – This is used to choose a specific cell when the computer wants to
read or write data.​

2.​ Control terminal – This tells the cell whether the operation is a read or a write.​

3.​ Data terminal – This is used in two ways:​

○​ During a write, it carries the signal (1 or 0) that will be stored in the cell.​

○​ During a read, it carries the output, which is the value (1 or 0) already stored in
the cell.
Even though the actual design inside a memory chip can be very complex and depends on the
technology used, the main idea stays the same: each memory cell can be selected, written into,
and read from — just like shown in Figure 5.1.

DRAM and SRAM

All the memory types in this chapter are called random access memory, which means you can
go directly to any part of the memory to read or write data — you don’t have to go through it in
order like a tape.

In Table 5.1, different kinds of semiconductor memory are listed. The most common type is
called RAM (short for Random Access Memory). Technically, the name "RAM" is not fully
correct, because all the memory types in the table are random access. But in everyday use,
RAM refers to memory that can both read and write data quickly and easily.
What Makes RAM Special?

●​ RAM uses electrical signals to read and write data.​

●​ One important thing to know: RAM is volatile. That means if the power goes out, all the
data stored in RAM disappears.​

●​ So, RAM is only used for temporary storage — like when your computer is running and
storing apps or files you're using right now.​

●​ The two main kinds of traditional RAM are:​

○​ DRAM (Dynamic RAM)​

○​ SRAM (Static RAM)​

●​ Some newer RAM types don’t lose data when the power goes off, but we’ll learn about
those in another section.

DRAM (Dynamic RAM)

DRAM stores each bit of data using a capacitor (a small part that holds an electrical charge).
The charge tells the computer if it’s a 1 or a 0.

●​ A problem with capacitors is they leak charge over time.​

●​ So DRAM has to be refreshed often, meaning the computer has to keep recharging it to
keep the data.​

●​ That’s why it’s called dynamic — the data is always changing and needs maintenance.

​
Look at Figure 5.2a — it shows a single DRAM memory cell:

●​ The address line picks the cell to read from or write to.​

●​ A transistor acts like a switch: it opens when voltage is applied (so electricity flows), and
closes when it’s off.​

●​ To write a bit:​

○​ A voltage signal (high for 1, low for 0) is sent to the bit line.​

○​ The address line is activated to store the charge into the capacitor.​

●​ To read a bit:​

○​ The transistor turns on, and the charge from the capacitor is sent to a sense
amplifier.​

○​ The amplifier checks the charge and decides if it's a 1 or 0.​

○​ Important: reading discharges the capacitor, so it must be refreshed again.​

DRAM is a bit tricky because although each cell stores just one bit (0 or 1), it works more like an
analog device — storing a range of charge levels, not just fully on or off.

SRAM (Static RAM)

SRAM works differently. Instead of using a capacitor, it uses a flip-flop circuit — a small digital
circuit made from logic gates (just like in a processor).

●​ SRAM will hold its data as long as power is on.​

●​ It does not need to be refreshed like DRAM.​

Look at Figure 5.2b — it shows an SRAM memory cell:

●​ It uses four transistors (T1, T2, T3, T4) arranged to keep a steady state of either 1 or 0.​

○​ If point C1 is high (1), then point C2 is low (0), and vice versa.​

●​ This setup is stable and doesn't lose its data — as long as the power stays on.​
 ●​ Two more transistors (T5 and T6) are used to read or write the value:​

○​ For writing, a value is sent on one line (say, 1), and the opposite on another line
(0), to force the circuit into the correct state.​

○​ For reading, the value is simply read from one of the lines.

SRAM vs. DRAM

Both SRAM and DRAM are volatile (they lose data when the power is off), but they work very
differently:

Feature DRAM SRAM

Stores data Capacitor Flip-flop (logic gates)

using

Needs Yes No
refreshing

Size Smaller cells Bigger cells

Cost Cheaper More expensive

Speed Slower Faster

Use Main memory Cache memory

●​ DRAM cells are smaller and cheaper, so they are used for main memory in
computers.​

●​ SRAM is faster and more stable, but also larger and more costly, so it’s used for
cache memory (which is smaller and used more often by the processor).
🔹 What is ROM?
ROM stands for Read-Only Memory. Just like the name says, it's memory that you can only
read — you can’t change the data once it’s stored (at least not easily).

●​ ROM is nonvolatile — it keeps data even when the power is off.​

●​ ROM is often used to store permanent instructions that help start a computer or run
basic functions.​

Some examples of where ROM is used:

●​ Microprogramming (small control programs)​

●​ Frequently used subroutines (library functions)​

●​ System programs​

●​ Function tables (like math tables used in calculations)​

📌 ROM is like a "permanent memory book" your computer can open and read but
can't rewrite.

🔹 How ROM Is Made

ROM is made during the chip's manufacturing process. The data is wired directly into the chip.

This has two main problems:

1.​ It's expensive to set up the data (even if you're making just one chip).​

2.​ If there's a mistake in the data, the whole batch of chips may have to be thrown away.​

So ROM is great if you’re making many copies, but not ideal for custom or small runs.

📊 Types of ROM
Let’s look at different types of ROM. We can imagine a table comparing them based on how
they’re written, erased, and used:
 Type Writable? How Erased? Reusable Notes
?

ROM ❌ No Not erasable ❌ No Programmed once at the factory

PROM ✅ Once Not erasable ❌ No Written once with special

equipment

EPROM ✅ Yes UV light (whole chip) ✅ Yes Can be erased and

reprogrammed

EEPRO ✅ Yes Electrically ✅ Yes Can erase/write small parts

M (byte-level)

Flash ✅ Yes Electrically ✅ Yes Fast and common in USB/SSDs

(block-level)

🔹 PROM (Programmable ROM)

PROM is like regular ROM, but it can be programmed once — after the chip is made.

●​ You use special equipment to burn data into it.​

●​ It’s useful when you want to store permanent data later, not during chip creation.​

●​ Still nonvolatile, and cannot be changed after writing.

🔹 EPROM (Erasable Programmable ROM)

EPROM can be erased and reprogrammed, but not easily.

●​ To erase it, you need to shine UV light through a small window on the chip.​

●​ This clears all data, and then you can write new data.​
 ●​ The erasing takes time — sometimes up to 20 minutes.​

🔍 Imagine a chip with a little “sunroof” — you shine UV light through it to clean the
data!

🔹 EEPROM (Electrically Erasable PROM)

EEPROM is more advanced.

●​ You can erase and rewrite data electronically — no light needed!​

●​ You can update one byte at a time, which gives more flexibility.​

●​ The writing is slower than reading, but it’s much easier to work with than EPROM.​

💡 EEPROM is great for things like storing BIOS settings in your computer.
🔹 Flash Memory
Flash is a very popular type of memory used in:

●​ USB drives​

●​ SSDs​

●​ Memory cards​

It works like EEPROM but:

●​ It erases blocks of memory, not bytes.​

●​ It’s faster and cheaper than EEPROM.​

●​ It’s very dense, meaning more data fits in a small space.​

🔥 Flash gets its name because it can erase memory in one quick "flash."