Cambridge AS & A Level · Part 1 · Chapter 3
Hardware
Where data lives, how it gets in and out, and why a computer needs so many different kinds of memory. Every figure, question and exam question from the chapter is here, with labs to try each idea.
Guru Jazzy🎯 Learning objectives
By the end of this chapter you should be able to:
- explain why a computer system needs input, output, primary memory and secondary storage, including removable storage
- explain what an embedded system is and how it uses hardware
- describe how the main hardware devices work
- Explain why buffers are used
- explain the difference between RAM and ROM
- explain the difference between SRAM and DRAM
- explain the difference between PROM, EPROM and EEPROM
🧭 3.01 What hardware has to do
A general-purpose computer system needs three abilities:
Talking about storage
| Word | Meaning |
|---|---|
| Memory | Components the processor can reach directly. |
| Primary storage | Another name for those memory components: main memory and cache. |
| Secondary storage / file-store | Where files are kept for the long term. The processor cannot use it directly; data has to come into main memory first. |
🪜 The memory hierarchy Figure 3.01
Move down the hierarchy and access time, capacity and physical size all increase. Move up and the cost per byte increases.
Everyone would like lots of cheap primary storage that is also fast. That is not possible: the fastest parts cost the most and hold the least. So every real system is a compromise. You could argue secondary storage exists because all-primary storage would cost too much, but it is simpler to say that keeping data for the long term needs its own dedicated components.
🔬 Lab: if one nanosecond were one second
Tap a level. The times are rough, typical values, scaled up so that 1 ns becomes 1 second.
🗄️ Where can storage sit?
Another way to choose a storage medium or device is to ask where it sits:
| Type | What it means | Examples |
|---|---|---|
| Integral | Built into the system; the user cannot normally get at it | Hard disk, solid-state drive (SSD) |
| Removable media | An item you insert into a drive in, or attached to, the computer | Floppy disk, optical disc, magnetic tape cartridge |
| Peripheral | A whole device you plug in | External hard drive, memory stick, memory card |
| Portable | Small enough to carry away, e.g. as a backup | Flash memory stick |
| Remote | Somewhere else, reached over a network | Cloud storage; a data centre's magnetic tape, RAID or storage area network (SAN) |
🎮 Sort the storage
Where does each one sit? Pick a category.
⇄ Getting data in and out
Ways to output data
- Screen display
- Hard copy from a printer or plotter
- A virtual reality headset
- Sound through a speaker
- Writing it to a storage device
- Sending it over a network
Ways to input data
- A keyboard or keypad
- the screen: icons, menus, a pointing device or touch
- A game controller
- A scanner
- a microphone, with voice recognition software
- Reading it from a storage device
- Receiving it over a network
🔌 3.02 Embedded systems
There are far more embedded systems in the world than general-purpose computers. In an embedded system the processor, memory and I/O are usually built into one chip, called a microcontroller.
Its I/O can be one of three kinds:
- internal only: the system just talks to the device it sits in
- monitoring and control: sensors in, actuators out (Chapter 11)
- a full user interface, as on a mobile phone
- Each one is special-purpose, built for a single job.
- They are made in huge numbers, so economies of scale make them cheap.
- Very little memory, so only short programs.
- To fix a faulty program, the whole chip had to be replaced.
Today many embedded systems are part of the Internet of Things (IoT). A network connection makes them much more useful, but it also makes security a real concern.
One possible answer
The appliance joins the home Wi-Fi. The microcontroller sends a small message, such as "wash finished" or "door left open", to the maker's cloud server. The server pushes a notification to an app on the owner's phone. Every hop needs securing: a weak password on the appliance gives an attacker a way into the home network.
🧠 3.03 Memory components: RAM and ROM
- RAM is random-access, also called direct-access: any location can be reached directly. It is read–write and volatile.
- ROM is also random-access, but it cannot be written to while it is in use in the system, and its contents survive power-off: it is non-volatile.
Two kinds of RAM
| DRAM (dynamic) | SRAM (static) | |
|---|---|---|
| Stores each bit in | a capacitor, which slowly leaks charge | a flip-flop (Chapter 19.02) |
| Refreshing | must be refreshed every few milliseconds | no refresh needed while powered |
| Components per bit | fewer | more |
| Cost and density | cheaper, more bits per chip | more expensive, fewer bits per chip |
| Access time | longer | shorter (faster) |
| Typical use | main memory | cache memory; embedded systems that need only a little RAM |
What ROM is for
ROM holds data or programs used unchanged, over and over. In a general-purpose computer its most important job is holding the bootstrap program, which runs the moment the system is switched on. ROM is also used in many embedded systems.
Four types of ROM, from least to most flexible
- ROM: the contents are installed when the chip is manufactured. New contents mean a new chip.
- PROM (programmable ROM): the chip maker supplies blank chips; the system builder writes the program or data. The builder can test a few chips before programming the whole batch. Once written, the contents cannot be changed.
- EPROM (erasable PROM): the contents can be erased with ultraviolet light and rewritten, but usually the chip has to be taken out of the circuit first.
- EEPROM (electrically erasable PROM): like EPROM, but erased by an electrical signal, so the chip can stay in the circuit while its contents are changed. In use it is still read-only.
🎮 Which memory would you choose?
Here is what this browser is allowed to tell us. For the full answer, look at Settings → System → About (Windows), About This Mac, or Settings → About phone.
🚦 Buffers
When data moves from one part of a system to another, the sender can often go faster than the receiver can cope. The answer is a buffer: data goes into it first and leaves for its destination later. A buffer works as a queue, so data comes out in the same order it went in. The buffer is usually an area of main memory.
🔬 Lab: a print buffer
The processor sends characters to a printer through an 8-slot buffer. Set the speeds and press Run. When the buffer is full, the sender has to wait.
Printed so far:
Show some answers
- Processor → printer (a print buffer or spooler).
- Keyboard → processor (the keyboard buffer keeps keystrokes typed while the computer is busy).
- Network → video player (streaming video is buffered so playback does not stutter).
- Main memory → hard disk when a file is saved.
- Sound card → speaker, and microphone → sound card.
💽 3.04 Secondary storage devices
🧲 Magnetic media
Magnetic tape was used to record sound long before computers existed, so it became the first computer storage device. The hard disk came a few years later and was invented specially for computers. Both store data by magnetisation.
- A read head uses the fact that a magnetised spot changes an electrical property as it passes.
- A write head does the reverse: an electrical current sets the magnetisation.
- The two are combined into one read–write head. The two possible states of magnetisation stand for 1 and 0.
How a hard disk is built
- There is more than one platter (disk), and each side has its own read–write head.
- All the platters spin together at the same speed.
- The heads sit on actuator arms that move them across the surfaces, and all the heads move together.
- A cushion of air keeps each head from touching its platter.
- Data sits in concentric tracks. Each track is divided into sectors holding a fixed number of bytes; a sector is the smallest unit of storage.
- Because the heads move together, the same track on every surface can be read without moving the arm. That set of tracks is a cylinder.
A hard disk is a direct-access read–write device, because any sector can be chosen. Inside a sector, though, the data is read sequentially, in order.
This is a simplified picture. For example, outer tracks are longer than inner ones, and designers have to allow for that or capacity is wasted.
🔬 Lab: find a sector
Tap any sector. The head moves to that track, then waits for the sector to spin round underneath it.
Fragmentation
A file needs enough sectors, but they do not have to be next to each other. As files are created, deleted and edited, their sectors get scattered (fragmented), so the head moves more and the disk slows down. A defragmentation program reorganises the sectors so each file is in one piece again (Chapter 8.03).
🔬 Lab: fragment, then defragment
Save and delete files a few times, watch them break up, then defragment.
💿 Optical media
Like tape, optical storage grew out of technology that had nothing to do with computers:
- CD-DA (digital audio) led to CD-ROM, widely used to distribute software, though it could not replace the floppy disk.
- CD-RW (read–write) finally made the CD a full alternative to floppy disks.
- The DVD (first "digital video disc", later "digital versatile disc") replaced the CD.
- The most powerful is now the Blu-ray disc (BD).
This drive reads a CD with infrared laser light (780 nm) or a DVD with red laser light (680 nm).
Reading the disc
- The disc has one spiral track from the centre to the outer edge.
- The disc spins while the laser moves across, staying focused on the spiral.
- The track has pits and lands, and the laser reflects off the surface.
- A pit reflects differently from a land; the detector picks up the change in light intensity.
- That difference is read as a 1 or a 0.
🔬 Lab: read the track
In this simplified model a land (bright) is read as 1 and a pit (dark) as 0. Scan the track, then type the 8 bits. (Real discs actually encode a 1 wherever pit changes to land or land to pit.)
Writing: CD-RW and DVD-RW
The reflective layer is a special alloy. When data is written (burned), the laser heats a spot until it melts. Depending on the laser's intensity, it then cools into either a crystalline solid, which reflects, or an amorphous solid, which does not. That difference codes 1 and 0.
Even with a single track, the disc is direct-access: the laser can jump forwards or backwards, and the track is split into sectors much like a hard disk's.
Why Blu-ray holds more
Capacity depends on how close together the bits can be packed. Rotation speed matters, but the wavelength of the light matters most: shorter wavelengths focus to a smaller spot.
| Disc | Laser | Wavelength | Typical capacity |
|---|---|---|---|
| CD | infrared | 780 nm | about 700 MB |
| DVD | red | 650–680 nm | 4.7 GB per layer |
| Blu-ray | blue-violet | 405 nm | 25 GB per layer |
⚡ Solid-state media
- Based on flash memory: semiconductor technology with no moving parts.
- The circuits are arrays of transistors acting as memory cells. The common type is called NAND because the circuit resembles a NAND gate (Chapter 4.04), with cells connected in series.
- A NAND flash controller handles reading and writing.
- A whole block of cells is erased at once, "in a flash". Before new data can be written to a block, the block must be erased first.
- A block holds several pages; one page can be read in a single operation.
🔬 Lab: erase before you write
Two blocks of 4 pages. Tap a page to write to it. An empty page can be written straight away. A used page cannot be overwritten, so the controller has to copy out the other pages, erase the whole block and write everything back. Watch the erase counts: every erase wears the cells a little.
Flash is used in memory cards and in USB flash drives (memory sticks), currently the first choice for removable storage. Newer technologies, such as phase-change RAM (PRAM), may one day replace it.
Used in place of a hard disk, it is a solid-state drive (SSD). With no moving parts you might expect it to last forever. It does not: the material slowly wears with use. The wear is gradual, though, and can be detected and corrected for. The big advantage over a hard disk is much faster access.
🔎 Extension Question 3.01
Research the storage technologies on sale today.
- Make a table of the storage options inside a laptop, showing cost, capacity and access speed for typical examples.
- Make a similar table for peripheral storage devices.
- Which technologies are still worth buying, and which are becoming uncompetitive? Are any new ones about to arrive?
A starter table: the capacities and speeds are rough typical figures. Tap a Price cell to fill in what you find in a shop today.
| Device | Where | Typical capacity | Typical speed | Price (you find it) |
|---|---|---|---|---|
| 2.5″ hard disk | Laptop | 1–2 TB | ~100–150 MB/s, ~10 ms access | |
| SATA SSD | Laptop | 0.5–4 TB | ~550 MB/s, ~0.1 ms access | |
| NVMe SSD | Laptop | 0.5–4 TB | ~3,000–7,000 MB/s | |
| USB flash drive | Peripheral | 32 GB–1 TB | ~100–400 MB/s | |
| External hard disk | Peripheral | 1–5 TB | ~100–150 MB/s | |
| Portable SSD | Peripheral | 0.5–4 TB | ~1,000–2,000 MB/s | |
| Memory card (SD) | Peripheral | 32 GB–1 TB | ~90–300 MB/s | |
| Blu-ray / DVD writer | Peripheral | 25–100 GB / 4.7 GB per disc | ~20–50 MB/s |
🖥️ 3.05 Output devices: the screen
Chapter 1.05 showed how a bitmap image is stored as pixels. A screen is built from pixels too, with one big difference: each screen pixel is made of three sub-pixels, usually red, green and blue. Changing how much light each sub-pixel gives out produces the full range of colours.
🔬 Lab: mix a pixel
From CRT to LCD
An old cathode ray tube (CRT) has no separate part for each pixel. The inside of the screen is coated with phosphor, which glows when electrons hit it; steering the electron beam lights up a pixel. Colour CRTs have red, green and blue phosphors arranged in an array.
Flat screens now dominate. In a liquid-crystal display (LCD) every pixel is a cell of liquid crystal:
- A backlight, usually LEDs, shines through a polariser, so polarised light reaches the pixel matrix.
- A second polariser sits between the pixel matrix and the screen.
- A voltage across a pixel cell changes how its liquid-crystal molecules line up.
- That changes the light's polarisation, so it changes how much light gets through the second polariser, and what you see.
🥽 Virtual reality headset
The key parts are the two eye-pieces. The system feeds each eye a slightly different image, and together they look like a 3D world. The images can come from special photography or a 3D graphics package. The wearer chooses what to look at by moving their head or using a controller.
🖨️ Hard copy: printers
Two technologies dominate printing, and both can print text or images.
Inkjet printer
A sheet is fed in. The printhead moves across, its nozzles spraying droplets of ink; the paper moves forward a little and the head crosses again, until the page is done. The ink comes from one or more cartridges.
Laser printer
🎮 Put the laser printer in order
Use ▲ and ▼ to put the nine stages for one black-and-white page in order, then check.
Colour: laser printers need a separate toner for each of cyan, magenta, yellow and black, and the process runs once per colour. Inkjets do the same with separate colour inks. Both print dots: quality depends on the dots per inch, and software controls how many dots make up each pixel.
📐 Hard-copy graphics: plotters and 3D printers
A graphic is stored as a bitmap or as a vector graphic (Chapter 1.05). Printers print bitmaps, so a vector file is converted to a bitmap first. Technical drawings often need more accuracy, so they use a graph plotter:
- pens draw on a large sheet of paper held by sprockets along two sides
- the sprockets move the paper forwards or backwards; each pen is either parked or drawing
- the circuitry and software draw straight from the original vector file
Engineers and designers in manufacturing use plotters. They may also use a 3D printer, another technology for computer-aided manufacture (CAM):
- Design the object in a computer-aided design (CAD) package.
- Software splits the design into layers.
- Each layer's data goes to the printer, whose nozzle squirts material onto the printer bed to build that layer.
- Repeat, layer by layer.
- Cure the finished object so the layers bond and the material reaches its final form.
🔬 Lab: print in layers
The bionic ear in Figure 3.06 used three "inks": silicone for the structure, a gel containing chondrocyte (cartilage) cells, and silicone filled with silver nanoparticles. It was cured by incubating it in a culture medium so the cells could grow cartilage. The only part missing was skin.
⌨️ 3.06 Input devices: the keyboard
A key press seems to put a character on the screen instantly, but that is an illusion. The press is turned into a character code and sent to the processor; the processor, under the operating system's control, displays the character. A shortcut key combination works the same way, except that the processor carries out an action instead.
The keyboard has its own circuitry, a microprocessor and a ROM chip:
- The keys sit above a key matrix: a set of row wires and a set of column wires.
- Pressing a key makes contact where its row and column cross.
- The microprocessor keeps testing every row/column circuit to see if one has closed.
- When one closes, it works out which intersection caused it.
- It looks up that intersection in the ROM to find the character code, and sends the code on.
🔬 Lab: the key matrix
Press a key, or type one of these on your own keyboard.
🎮 Be the microprocessor
🖱️ The screen as an input device
Early systems had only a keyboard and a monitor. Even then, software could show a menu and the user typed the number of an option.
Graphical user interfaces (GUIs) became standard on microcomputers in the 1980s. Their icons let the user control input, but they need a pointing device such as a mouse to move a cursor. With a mouse click, the screen became an input device as well as an output device.
👆 Touch screens
Early touch screens worked on CRTs but suit any screen. Emitters along the edges sent infrared light or ultrasonic waves to detectors on the opposite side. A finger blocked some of the beams, and those detectors measured a weaker signal.
Flat screens allowed new methods: touch-detecting layers just below the surface.
The two layers have a thin gap between them. Pressing bends the top layer onto the lower one, and the point of contact forms a voltage divider horizontally and vertically.
The best type is projective capacitive touch (PCT) with mutual capacitance: an array of capacitors under the glass. It can detect several fingertips at once.
Whatever the type, the processor takes readings from the measuring devices, calculates where the touch was, and then carries out what the user asked for.
🎮 Which touch technology?
| Technique | Advantages | Disadvantages |
|---|---|---|
| Keyboard + numbered menu | Works on any screen; quick for experts | Slow to learn; few options per screen |
| Mouse / touchpad + GUI | Precise pointing; easy to learn | Needs a flat surface or space; slow for repeated tasks |
| Infrared / ultrasonic grid | Fits any screen; nothing in front of the image | Low resolution; dirt or stray objects cause false touches |
| Resistive touch | Cheap; works with a gloved finger or any stylus | Needs pressure; soft surface scratches; usually one touch only |
| Capacitive (PCT) touch | Multi-touch gestures; tough glass; very responsive | Does not work with ordinary gloves; costs more |
| Stylus / pen | Accurate drawing and handwriting | Easy to lose; needs a compatible screen |
📷 Input of a graphic
- A webcam streams video into the computer.
- A digital camera stores photos and videos, which are downloaded to the computer.
- A scanner reverses printing: it makes a digital copy of an image, which may be text. The sheet stays still while a light source the width of the paper moves from one end to the other. Mirrors and lenses send the reflected light to a charge-coupled device (CCD).
Three things to know about a CCD:
- It is an array of photo-sensitive cells
- each cell gives an electrical response proportional to the light intensity
- an analogue-to-digital converter turns these responses into digital values for the computer
🎙️ 3.07 Input and output of sound
IP telephony and video conferencing need voice in and voice out. Voice recognition is another way to input data, and voice synthesis is used more and more.
Input: microphone
- A flexible diaphragm vibrates when sound reaches it.
- Circuitry turns the vibration into a changing electrical signal. A condenser microphone uses a change in capacitance; another kind uses a piezoelectric crystal.
- An ADC (analogue-to-digital converter) turns the analogue signal into digital data the computer can process.
Output: speaker
- The reverse process: a DAC (digital-to-analogue converter) turns digital data into an analogue signal.
- The signal is a varying current through a coil inside a permanent magnet's field.
- As the current changes, the coil moves back and forth, moving a diaphragm that makes the sound.
Both are controlled by the sound (audio) card. Music is recorded or streamed the same way as voice. Some recorders convert to digital very early, so all the processing is digital. Music can then be played through speakers or stored for later.
🔬 Lab: inside the ADC
The grey curve is the analogue signal. The ADC samples it at regular times and rounds each sample to one of 2bits levels. The green steps are what the computer stores.
📌 Summary
- Primary storage is main memory: RAM (DRAM or SRAM) and ROM (possibly PROM, EPROM or EEPROM).
- Secondary storage includes magnetic, optical and solid-state media.
- Output devices include screens, printers, plotters and speakers.
- Input devices include the keyboard, scanner and microphone.
- Screens can be used for both input and output.
🃏 Key-term flashcards
Tap a card to flip it.
✅ Quick quiz
📝 Exam-style questions
Write your answer first, then open the model answer. These are my mark-scheme-style points, not official mark schemes.
(i) Any two, each explained (2 marks each):
- RAM is volatile, so its contents are lost when the power goes off; ROM is non-volatile and keeps them.
- RAM can be written to as well as read; ROM can only be read while in use.
- RAM holds the programs and data in use now; ROM holds fixed programs such as the bootstrap.
(ii) Both are random (direct) access. Or: both are primary storage the processor can reach directly.
(iii) DRAM stores each bit as charge on a capacitor, which leaks and must be refreshed. SRAM uses flip-flops and needs no refresh, so it is faster but costs more per bit.
(i)
- Magnetic: read–write heads detect or set magnetisation on spinning platters.
- Optical: a laser is reflected off pits and lands on a single spiral track.
- Solid state: no moving parts; NAND flash cells are erased a block at a time before being rewritten.
(ii) Magnetic and optical: both spin the medium, and both split tracks into sectors, which gives direct access. Or the two disc types read 1s and 0s from two different physical states of the surface.
(i) Any four:
- The keys sit above a matrix of row and column wires.
- Pressing a key closes the circuit at one intersection.
- The keyboard's microprocessor keeps scanning for a closed circuit.
- It identifies which intersection is closed.
- It looks up that intersection in ROM to get the character code.
- The code is sent to the computer's processor, which displays the character.
(ii) Voice recognition: the user speaks into a microphone, an ADC digitises the sound, and software matches it to words. Or: tap letters on an on-screen keyboard on a touch screen, or scan a printed page and use OCR.
(i) The drum is given an electric charge → the drum turns step by step → at each step the laser, steered by mirrors and lenses, discharges selected positions across the drum (or leaves them charged).
(ii) Any two: charged toner is applied and sticks only to discharged positions; the drum rolls over charged paper and transfers the toner; the paper is discharged and passed through heated rollers that fuse the toner.
(iii) There is a separate toner for each colour (cyan, magenta, yellow, black), and the process is repeated for each one.
- Emitters are fixed along the sides of the screen…
- …with detectors on the opposite sides.
- They send infrared light or ultrasonic waves across the screen.
- A finger blocks some beams, so some detectors get a weaker signal.
- The processor works out the touch position from which detectors are affected.
Resistive: two resistive layers sit under a flexible surface with a small gap between them. A press pushes the top layer onto the bottom one. The contact point forms voltage dividers horizontally and vertically, and the processor calculates the position from the voltages.
Or capacitive (PCT): an array of capacitors sits under rigid glass. A finger changes the capacitance near it. The processor finds the position from the changes and can detect several touches at once.
- Hard disk: magnetic: spinning platters coated with a magnetisable material.
- DVD-RW: optical: a disc with a phase-change alloy layer that is read, and written, by laser.
- Flash memory: solid state: semiconductor (NAND) memory cells with no moving parts.
- The disc spins; data is on one spiral track from the centre outwards.
- A red laser follows the track and is focused onto it.
- Writing: the laser heats the alloy until it melts. Depending on the laser's power, it cools into a crystalline (reflective) or amorphous (non-reflective) state.
- Reading: a lower-power laser is reflected from crystalline areas but not from amorphous ones. The detector reads the change in intensity as 1s and 0s.
- The track is split into sectors and the laser can move either way, so access is direct.
Cambridge International AS & A Level Computer Science 9608, Paper 12 Q1, November 2015
- RAM is volatile; ROM is non-volatile.
- RAM can be read and written; ROM can only be read.
- DRAM must be refreshed regularly; SRAM does not need refreshing.
- DRAM uses capacitors; SRAM uses flip-flops.
- SRAM is faster (shorter access time).
- DRAM is cheaper per bit, with more bits per chip.
- DRAM is used for main memory; SRAM for cache.
Cambridge International AS & A Level Computer Science 9608, Paper 12 Q6, November 2016
Guru Jazzy · Cambridge AS & A Level Computer Science · Chapter 3 Hardware · Figures from the Cambridge coursebook (Langfield & Duddell)