By proceeding to accessing this website, you unconditionally agree to its terms and conditions. If not, please stop and leave now.

Site Logo

‘PC Hardware Explained’ book
:: The illustrated guide to personal computer components ::
:: Best laptop/desktop buyer guide for 2022 ::

Cover of the book ‘PC Hardware Explained’

Modern computers are based on a modular design. This enables anyone with a screwdriver set to assemble a fully functional desktop computer that works just as good as a store‑bought branded PC. Assembling is easy. Knowing which computer components will work together can be a daunting task for a newbie. This book solves that problem. It explains computer hardware using simple terms, illustrations, photographs and tables. The reader will be able to read the technical specifications of a PC and understand what it can and cannot do. The mumbo‑jumbo accompanying the sales pitch of a new computer will not seem so alien. Whether assembling a new computer from parts or buying a new desktop/laptop from the store, this book will help making the correct choice. (There is a chapter with generic steps for assembling a computer.) After this 𝓕𝓤𝓛𝓛-𝓒🍅𝓛🍅𝓤𝓡 𝓟𝓐𝓟𝓔𝓡𝓑𝓐𝓒𝓚 has been read, it can serve as a ready‑reference for hardware terminology.

Kindle ebookAmazon₹70$3

Subscribe to my email newsletter

eBook cover

The details are mentioned in my author newsletter blogpost. Full/sample PDF ebooks of several books are listed in it. You do not have to subscribe to download them. However, if you do subscribe, you get a FREE 200-puzzle combined ebook version of my World of Word Ladders books.

Word Ladder sample

The old ‘PC Hardware Explained’ article

‘PC Hardware Explained’ lasted more than 20 years as an article ­­- the oldest article on this site. It was the only one of its kind on the Internet for several years before others copied the idea. Now, there are countless pretenders.

Google search for *pc hardware explained*

The old article will continue to be available for whoever who needs it.

Because of the modular nature of computer components, anyone with a Philips screwdriver can assemble a personal computer (PC) in under half an hour. However, the challenge for a layman is in knowing what each component looks like and identifying which ones will work together.

In this writeup, you will learn about the various components that make up a PC. You will be aided by photographs or diagrams so you have a sure idea what you are reading about. You will also learn some "computer hardware lingo" so that you understand what is mentioned in computer advertisements, user manuals, and support information. At best, you should be able to handle basic PC troubleshooting, maintenance and repairs all by yourself.


The processor is the main part of the computer. You might say it is the brain or the heart of a computer. Today's PCs use CPUs made by just three companies - Intel, AMD, and Via.

Processors come with technical and marketing mumbo-jumbo that you will have to be familiar with. Here are some.


INTERNAL CLOCK SPEED: CPUs or processors are often graded by their internal clock speed. This speed is the number of operations a processor can perform per second. For example, the Pentium-4 3.0 GHz has a clock speed of 3000 KHz or 3.0 GHz.

EXTERNAL CLOCK SPEED: This is the speed with which the CPU can communicate the rest of the system. The CPU communicates with the rest of the system via a gateway called system bus or front side bus. Hence, the external clock speed is also called Front Side Bus (FSB) speed.

MEMORY CLOCK SPEED: This is the speed with which the RAM is given requests for data.

In old CPUs, the memory clock speed used to be the same as the FSB. With the Athlon K7 processor, AMD introduced something new. The Athlon processor had an FSB of 100 MHz but the bus could make 2 data fetches per cycle from the RAM. So, the effective rate became 200 MHz. AMD then went on to claim that the Athlon had an FSB of 200 MHz. AMD called this Double Data Rate (DDR) FSB.

Intel later introduced this feature in their first Pentium-4 processor. In the P-4 processor, the system clock operated at 100 MHz and the bus was sampled 4 times, which resulted in the FSB being touted as 400 MHz (100 x4). Intel called this Quad-Pumped Bus. As of December 2006, external frequencies were at 200 MHz and Intel has been touting a FSB speed of 800 MHz for their processors.

CPUExterna Clock Speed or True FSBNumber of times System Bus is sampledMemory Clock Speed orAdvertised FSB (MHz)
Pentium II
AMD K71002200
Pentium III (Coppermine)1334533
Athlon (Thunderbird) "C"
Athlon XP (Thoroughbred)
Athlon XP (Barton)1662333
Pentium-4 (Northwood)2004800
Athlon FX (Sledgehammer)
Athlon 64 (Clawhammer)

Both Intel and AMD advertised their memory clock speed as the FSB of their processors. The correct FSB measure would have been the external processor frequency, as few components other than the RAM can operate at such high frequencies.

AMD later introduced a technology called HyperTransport, which integrated the memory controller in the CPU itself, thus eliminating the Front Side Bus (FSB)! This is a radical change in PC architecture because now the CPU could directly handle RAM operations and was not limited by an FSB. The PC architecture was also significantly changed with the HyperTransport Graphics Tunnel replacing the North Bridge and the HyperTransport I/O Bus Tunnel replacing the South Bridge. See the graphic below. With HyperTransport, all data is sent in packets instead of the earlier inefficient scheme of reserving bandwidth streams. HyperTransport's optimum use of bandwidth results in higher throughputs. Because of the parallel nature of HyperTransport, it initially operated at 1 GHz. Now, it works at a speed of 2 GHz. Despite the obvious advantage of HyperTransport technology, Intel continues to use its dated FSB-based architecture and makes inefficient CPUs.


Processors come in different shapes and fit into the motherboard through an electrical connection called the interface. There were two types of them - the slot and the socket. In the slot interface (used in old CPUs such as the Pentium II), the processor sat on the motherboard on one of its edges. When you look at it, it is as though the processor is sleeping on its sides on the motherboard. The socket interface, on the other hand, has the processor on all fours giving it a the-back-to-the-bed look. The socket interface is now standard and the slot interface is all but gone.

The CPU is fixed on the motherboard by placing it on the socket. See the picture for the CPU socket.1

Socket 7Pentium, AMD K6-2, K6-3
Super 7AMD K6-3
Slot 1Pentium II, Early Celerons
Slot AEarly Athlons
Socket 370Celeron, Pentium-!!!, C3
Socket 423Early Pentium-4s (Williamette)
Socket 478Pentium-4s
Socket A (Socket 462)Duron, Athlon, Athlon XP
Socket 754
(Supported Single Channel DDR-RAM)
Athlon 64 FX
Socket 940
(Supported Dual Channel DDR-RAM)
Athlon 64 FX
Socket 939
(Supports unbuffered Dual Channel DDR-RAM)
Athlon64 FX
Socket 940 (Socket AM2)
(supports DDR2 RAM)
Athlon64 FX, Athlon64 X2
LGA 775 (Socket 775)
Intel Core 2 Duo, Intel Core 2 Quad


The motherboard is the second-most important part of the computer. (See the sample motherboard layout at the top of the page.) It is the framework on which all other components literally sit on.

The motherboard has a brain of its own and this motherboard logic is called the chipset. A CPU needs to be paired with a motherboard that has a \ compatible chipset. A CPU may have many compatible chipsets made by different manufacturers.

Intel 440-BX2Pentium-II100100PC-66, PC-100 and PC-133 SD-RAM
Intel 815Celeron, Pentium-!!!66 & 13360 & 133PC-66, PC-100 and PC-133 SD-RAM
Via KT-133Duron, early Athlons100200PC-100, PC-133 SD-RAM
Via KT-266Later Athlons and Athlon XPs133266DDR 266
Intel 850Old Pentium-4s100400RDRAM
Intel D850Pentium-4 (Tualatin)133533RDRAM
Via KT-333Athlon )166333DDR 333
Via KT-400Athlon XP (Barton)200400DDR 400
NVIDIA nForce 430Athlon 64 Socket 939200HyperTransport 1 GHzDDR 400
Via K8T890Athlon 64 AM2 Socket200HyperTransport 2 GHzDDR2 800/667/533

The chipset of a motherboard is made to support a particular type or range of CPUs. So, if you choose to use a motherboard with particular chipset, then you have to pair it with a compatible CPU. The chipset also determines the choice of RAM you can use. Some chipsets support DDR SDRAM while most new ones use DDR2 SDRAM.

Apart from this, motherboard needs to provide support for a range of other requirements such as

Motherboard manufacturers woo customers by adding certain extra features. Many motherboards provide features such as onboard sound, video, Firewire, RAID, and Gigabit LAN, which typicaly used to require an add-on PCI card. Gaming enthusiasts are being wooed by overclocking options, which allow the system parameters to be tweaked above their rated levels, and support for dual or quad grahpics card setups.

Buyers should ensure that the layout of the motherboard components do not hinder the installation or the operation of other components. Advertisements typically carry the name of the chipset and the onboard graphics subsystem. For example, the Dell Dimension E521 has an nVIDIA nForce 430 chipset and an integrated nVIDIA GeForce 6150 LE graphics subsystem.

Motherboard chipsets are manufactured Acer, Intel, ATI, nVidia, and Via. Motherboards are manufactured by companies such as Asus, Gigabyte, MSI, Jetway, and Mercury.

System Memory

RAM or Memory

The CPU needs to temporarily store and retrieve large amounts of data during its operation. The amount of data that can be stored on the CPU itself (in CPU registers and CPU L1, L2, L3 caches) is very limited. So, the PC architecture uses RAM modules installed on the motherboard to provide a high-speed data buffer for CPU.

Today, there are two types of RAM that are being used - DDR RAM (DDR SDRAM) and DDR2 RAM (DDR2 SDRAM). DDR RAM (Double Data Rate Synchronous Dynamic RAM) is the older standard and is getting replaced by DDR2 RAM.

SDRAM was the predecessor of DDR RAM and was used by older CPUs such as Intel Pentium-IIIs and early AMD Durons and Athlons. SDRAM stores data in memory cells, which are activated by a clock signal. Data in the memory cells is fetched through a path in the RAM module called the external data bus. DDR-RAM differed from SDRAM because it could fetch twice the amount of data as SDRAM during the same signal cycle - once during the rising phase and once during the falling phase of the signal curve. DDR2 RAM had a data bus that works twice as fast as the memory cells (their clock speed) because it could fetch data from two different memory cells during the same signal cycle. Because the data from different cells need to be synchronised, it had significant latency (delay) issues initially. When DDR2 RAM which operated at higher speeds came out, the latency issue became less of a problem.

RAM operations are managed by a component called memory controller. On AMD CPUs, the memory controller is integrated on the CPU. For Intel CPUs, the memory controller is part of the motherboard. The memory controller is connected to the memory banks by one or two channels. Memory modules can be organised in a single-channel configuration or dual-channel configuration. In a single-channel configuration, any combination of compatible memory modules can be used together but they will use a single channel to the memory controller. In a dual-channel configuration, two identical sets of memory modules are used to employ two separate channels to the memory controller, effectively doubling the memory bandwidth.

Memory modules with different clock speeds can be used together but they will function at the lowest supported speed. Also, SDRAM, DDR RAM and DDR2 RAM cannot be used together. It is always best to buy all memory banks together with same specifications. When you add an extra module as a later upgrade, ensure you buy modules with equivalent (or, if there is no choice, higher) ratings.

The data transfer rate for DDR RAM is twice the memory clock speed because it performs two data fetches per clock cycle. For DDR2 RAM, it is four times the memory clock speed because it fetches twice the data during the same clock cycle as DDR RAM. The peak bandwidth of DDR2 RAM is calculated by the width of the data channel between the memory (RAM) and memory controller. Today, memory channels are 64 bits wide. With dual-channel memory configuration, this width is effectively doubled to 128 bits.

Peak bandwidth = Memory Speed x Number of Bytes Per Channel x Number of Channels

So for DDR-800 memory, it will be 800 x 64 x 1 equaling 6.4 GBs for single channel configuration.

DDR2 TypeMemory Clock SpeedData Bus SpeedData Transfer Per SecondPeak Bandwidth
(Single-Channel configuration)
DDR2-400100 MHz200 MHz400 Million3.2 GBps
DDR2-533133 MHz266 MHz533 Million4.267 GBps
DDR2-667166 MHz333 MHz667 Million5.333 GBps
DDR2-800200 MHz400 MHz800 Million6.4 GBps

Memory modules are manufactured by Samsung and Hynix and are available under brands Corsair, Kingston, Transcend and others. Instead of buying a single memory bank, it is best to buy two banks each with half the capacity and use them in dual-channel configuration. Refer the motherboard manual for instructions on how to use memory banks in dual-channel configuration.

Graphics Subsystem

Originally, graphics (video) function of a PC was provided by an ordinary PCI video card. Drawing graphics was not a big task and the CPU could handle the load. For intensive applications such as games, the need for a separate graphics subsystem arose and the Accelerated Graphics Port (AGP) was created. The AGP looked like a PCI slot but was used by a special AGP video card. Several iterations of the AGP standard came from AGP 1x to 2x to 4x to 8x. Later, the AGP standard was replaced by PCI Express (PCIe or PCI-E), which is currently universal. Unlike AGP and PCI standard, which used data channels in an exclusivel manner, PCIe standard allowed shared data channels among devices, which could then be grouped or limited dynamically to provide optimum bandwidth. The PCIe standard allowed for a maximum of 32 channels or lanes. A slow device could be given just one lane while a fast device could be theoretically given up to a maximum of 32 lanes.

Most motherboards use 16 lanes for the graphics subsystem. So, what was originally the AGP port is now replaced by a PCIe x16 slot.

nVidia introduced a technology called SLI (Scalable Link Interface) which allowed two PCIe x8 slots to be used by two graphics cards. For gamers, this resulted in great performance improvements because the graphics was rendered by two graphics card working together, instead of just one. ATI, later acquired by AMD, introduced a similar technology called CrossFire, which allowed the use of two ATI graphics cards on two PCIe x8 slots on motherboards using ATI chipsets. PCs using SLI or CrossFire graphic subsystems have significantly higher power requirements and need good cooling setups.

The graphics card if fixed on a PCIe x16 slot on the motherboard and its video-out port will be available on the rear of the cabinet. Older graphics cards have a D-Sub connector that can be used to connect to an analog cable of the CRT monitor. Newer graphics card have a DVI connector that can be used to connect to a digital cable to an LCD screen. However, many LCD monitors only provide D-Sub connections. So make sure your LCD monitor has a DVI connector. Almost all graphics card these days have DVI connectors only.

Apart from this connector, some video cards pack a TV tuner on board. This allows you to connect a CATV (Cable Access TV) cable to the video card and watch television programs. If the video card has a TV tuner, it most likely also has a video-in ports also, which allows you to capture or view video from a video output device such as a VCR or DVD player. Video-In ports could be either serial or composite (RCA) video connectors. (To capture/listen the audio part, you need to connect the audio output to the line-in jack of the sound card.) These cards are also known as VIVO cards. ATI's All-In-Wonder cards are the best VIVO cards on the market.

A company called Aegia has introduced a physics card called PhysX that relieves the graphics subsystem of all the physics-related calculation necessary to render graphics. The Aegia PhysX card comes in the form of a PCI add on card.

Disc Drives

Hard Disk

Hard disks, CD drives and DVD drives are known as IDE (Integrated Drive Electronics) devices. IDE devices are used to store data. Hard disks come in capacities of 20 to 750 GB. 1 GB = 1,000,000,000 bytes. Because PCs treat thousands as 1024, a 40 GB hard disk might be identified as having 37 GB = 40,000,000,000/1024 x 1024 x 1024. Good hard disks are available from Hitachi, Maxtor, Samsung, Seagate, and Western Digital.

Somewhat similar to the music stored on an audio tape, data is stored on a hard disc magnetically. The hard disk has one or two magnetically sensitive platters inside. When the hard disk is working, the platters are spun at a very high speed. The updraft from this motion allows a read-write head assembly to levitate over the platters. The drive mechanism moves the head assembly in the required positions to read or write data. When you switch off the PC abruptly or when the power goes off without warning, the head assembly does not get time to park itself safely away and instead crashes onto the platters. In serious cases, this renders the disk unusable - a crashed hard disk. Stellar Infomation Systems provides data recovery services and software for crashed hard disks.


Another measure for hard disk performance is the rpm, which is the speed at which the disk platters spin. These come in 5400 rpm for Ultra DMA/33 and Ultra DMA/66, 7200 rpm usually for Ultra DMA/100, and 10,000 rpm for SCSI hard disks. The hard disks also have some onboard memory called the cache. The size of this cache also contributes to its performance. The bigger the better.

There are three types of hard discs - parallel ATA (PATA) and serial ATA (SATA). PATA is the oldest technology and is being slowly being replaced by SATA. Older motherboards that support only PATA natively can be made to support by using a SATA add-on card. However, most new motherboards natively support SATA. SATA supports data transfers up to a theoretical 300 MBps. SATA2 capable hard disks are manufactured by Samsun, Hitachi, Seagate, Western Digital, and Maxtor.

Left: eSATA bracket for hot-plugging external hard disks Right: SATA and PATA connectors on the motherboard

External or portable hard disks are now in fashion and SATA-II technology incidentally allows for hotplugging of hard disks. An eSATA bracket is fixed to the rear of the cabinet. The cable on the bracket is connected to one of the SATA connectors on the motherboard. A user can then use a cable that came with the portable hard disk to connect the disk to the eSATA bracket.

IDE devices and IDE cables


IDE stands for Integrated Drive Electronics. Examples of IDE devices are the hard disk drives, CD drives, DVD drives, etc. These drives are always mounted on a set of racks on one side of the cabinet and connected to the motherboard on the IDE connectors using IDE cables. Most motherboards have at least two IDE connectors.3.

Each IDE connector can be connected to two IDE devices. If there are two IDE connectors in the motherboard, you can connect four IDE devices. If there are four IDE connectors, you can connect eight IDE devices.

Most people have only one hard disk and one CD drive. In such a situation, both IDE connectors can be used for connecting them. This lets each device to take over one entire IDE channel for itself.


When there are two devices that need to be connected to a single IDE connector or when there are more than two IDE devices, then one of the devices will have to be configured as Master and the other as Slave. Usually, the faster drive is configured as master. Otherwise, the slower drive will slow down the faster one.

Usually, though not always, the IDE cable has 3 connectors on its entire length. The connector at the longer end is connected to the motherboard. The other two connectors are connected to your IDE drives such as the hard disk and the CD or DVD drives.

Master/slave setting in IDE devices is done by closing the appropriate jumpers with a jumper cap.

The drives themselves have to be configured for their master/slave roles by using certain jumpers on IDE drives themselves. The first jumper from the right must be closed with a jumper cap if the drive is to be set as master. Close the second jumper from the right to make the drive slave. Please refer your manuals for the correct setting.

RAID (Redundant Array of Inexpensive Disks)

Many motherboards, these days, come with extra onboard IDE controllers. You may just connect additional IDE devices to them or you can implement a RAID pair on them. See the motherboard picture for the optional RAID connector. With RAID 0, you can mimick two identical hard disks as a single hard disk. This way, read and write operations would be distributed over 2 hard disks resulting in improved performance. To implement RAID-0, you have to attach preferably two identical hard disks (same model, brand, and capacity) to the connectors using two 80-pin cables. You may have to do some tweaking with drivers and BIOS updates before you can then treat the RAID pair as a single hard disk. Since the data being written is split between two hard disks, data in one of the hard disks is useless without the other. If one of them fails, all data is lost. RAID-0 should be attempted only if it is supported by the PC manufacturer. Most users use the extra IDE connectors for expansion rather than for RAID-0.

While RAID-0 is known as data striping, RAID-1 is known as data mirroring. In RAID-1, data being written on one hard disk is simultaneously copied to another hard disk. If one hard disk suffers a mechanical failure, your data will remain intact on the other hard disk.

RAID levels range from 0 through 5. There is one another level of RAID known as RAID 10, which is a combination of RAID 1 and RAID 0, and it is the best in terms of performance and redundancy.

Optical Drives and Disks

CD (compact disk) drives are used to read data recorded on a CD-ROM (compact disk read-only memory) disks. Data is recorded on CDs in the form of microscopic pits. The presence or absence of the pits, which is read by a red laser beam inside the drive, is used to represent digital data - the 1s and 0s. Factory-made CDs, such as audio CDs and video CDs, are made by pressing moulds, which create the pits on the disks.

On a CD-RW drive (CD writer), the pits are burnt on to a CD-R (CD writable) or a CD-RW (CD rewritable) disk using a more powerful laser. With a CD-R disk, data once written cannot be erased. On a CD-RW (CD rewritable), data can be written and erased countless times. A CD drive cannot read a DVD (Digital Versatile Disc) because the pits on a DVD are much more densely packed. To read a DVD, you need a DVD drive. DVD drives can also read CD disks. CD-RW combo drives have the ability to read DVD disks. To create or "burn" DVDs however, you need DVD-RW (DVD writers) drives. Liteon and Plextor make good CD and DVD writers. A CD writer or a DVD writer will not function as a writer until a CD/DVD burning application such as Nero Burning ROM is installed. Liteon ships a limited functionality Nero Express with its drives while some others provide the full version. CD-R and CD-RW disks come in capacities of 650 and 700 MB. Mini CDs, which are smaller in size and can be easily carried in a shirt pocket, hold 170 MB. Some CDs are available in the shape of a visiting card and are typically used to provide a multimedia version of a company or a business. DVD-R can store 4.7 GB of data. Dual layer DVD-Rs disk store data on two diferent lawyers and have a capacity of 8.4 GB. Double-sided DVD-R/DVD-RW disks store data on both sides of the disk and can store 17 GB of data.

Factory-made CDs have labels painted on them. However, with burnt CD consumables such as CD-Rs, CD-RWs, DVD-Rs DVD-RWs having neat labels is an issue. Some CD consumables come with no paint work and allow a sticky paper label. However, in high-speed drives, these labels might come unstuck and destroy the disk and the drive. Many people prefer using CD labeling pens. These pens have very soft tips and make permanent labels. (Using hard tipped pens scratches the disk.) Verbatim has introduced a new CD writer and disk technology called LightScribe. After writing the data, you flip the CD to burn the label using the writer. Both Lightscribe drives and Lightscribe disks can be expensive. Printers from Canon and Epson allow you to print on special printable CDs. If you need to print labels on plain-vannilla CD/DVD disks, get a CD printer from Techcom costing Rs. 4000.

Sony in association with other manufacturers and some Hollywood studios has introduced a new optical disc called the Blue Ray disc, which can hold more data. It is in competition with another optical disc standard known as HD-DVD promoted by Microsoft. Only future will tell which will become the dominant optical disc technology and replace the current DVD technology.

Expansion Slots

Peripheral Component Interconnect or PCI slots 7 are slots in which you can put addon expansion cards. The expansion card may be a sound card or a TV tuner or a network card or anything else that complies with the PCI standard.

Most motherboards have 3 to 5 PCI slots. Some have 6 PCI slots while some have just 1. Motherboards with integrated components usually have lesser number of PCI slots as the functionalities provided by addon cards are already on the motherboard.


PCI cards need something called IRQ (Interrupt Request Que) to operate, though some do not. A discussion of what IRQs are is unnecessary, but you may just be aware that a motherboard needs a specific number of IRQs (usually 16). Usually, most of these are already in use leaving 6 or so IRQs free for expansion cards.

Some PCI cards need an IRQ exclusively. If two such expansion cards start using the same IRQ, problems will arise. Some PCI slots share IRQs themselves. So, devices that need exclusive IRQs should not be used in them.

Windows 9x/Me need IRQs assigned in the BIOS, while Windows 2000 assigns IRQs dynamically.


When there is an IRQ conflict, you should try reassigning IRQs in the BIOS manually or try the card in a different PCI slot. By providing a device with a free IRQ, the problem will be solved. To know about IRQ assignments on your machine, try Start » Accessories » System tools » System Information. Here, click on Hardware Resources.

Refer your motherboard manual to know about interrupt assignments for your motherboard. Any interrupt whose function is marked as IRQ Holder for PCI Steering is a free IRQ. If you have a device that finds itself without an IRQ or having conflict with its current assignment, then change it to such an IRQ. Your problems will be solved.

Prior to this, go to the motherboard manufacturer's website and check for upgrades for the BIOS and chipset drivers. Also try to get the latest drivers for your PCI devices. After your upgrades, see if the IRQ problems are resolved. If the problems persist after the upgrades, try the manual IRQ allocation route.

IRQ problems can prevent a PC from being properly shut down or restarted. Many of those hanging-when-shutdown problems are due to this. Devices that take up a large amount of system resources should be placed higher on the PCI slots. A TV tuner card for example should be placed in the first PCI slot, while a sound card can be placed in the second or third PCI slot. When you install drivers for these devices, install them in the same order too.


PS2 keyboard port, PS2 mouse port, Parallel port, Gigabit Ethernet LAN port, Audio Channel ports, Optical S/PDIF audio port, USB ports, Composite video (RCA) port, Firewire port, and Serial port

In the early days of the PC, there were only serial and comm ports. Serial ports were used by mice and parallel ports were used by printers.

These days you have PS/2 and USB ports. PS/2 ports are used by keyboards and mice while USB ports are used by almost everything from printers to web cameras.

USB (Universal Serial Bus) devices, such as USB printers, were originally designed to daisy-chain devices - a single USB device is connected to the PC and all other devices are connected together as in a chain. However, manufacturers of USB devices stopped provide extra USB ports and motherboard started offerring more than two USB ports. In any case, USB devices could be hot-plugged - added or removed when the PC is on - so there was no need for providing the daisy-chaining functionality. USB ports today are typically used to provide connectivity to printers, broadband modems, USB data drive (thumb drives, memory drives), MP3 players (iPod), and mobile phones (Motorola V3 RAZR). USB technology is in its second revision with USB 2.0, which provides 480 MBps through put unlike the slow USB 1.1. Firewire ports (IEEE 1394), which came before USB 2.0, likewise are speed monsters and are now getting more common. The introduction of USB 2.0 has caused the Firewire to be somewhat overshadowed because USB 2.0 has a thoroughput of 480 mpbs while Firewire tops out at 400 mpbs. Firewire's second version is now in the making and is expected to have have a thoroughput of 800 mbps. Firewire's advantage over USB 2.0 is that it does not need a PC in between to do a data transfer between two Firewire devices.


Many people erroneously call the cabinet as the "CPU." The cabinet is basically a steel, aluminium or reinforced polymer box that holds the motherboard and other computer parts inside it. It has a power supply which provides the connection to the mains. Cabinets usually come in two shapes - vertical tower cabinets and the horizontal ones (usually seen in OEM PCs). Motherboards come in two different sizes or form factors to accommodate these cabinets.

ATX is a form factor and cabinet power supply standard. One must buy cabinets that conform to this standard. Its features allows the operating system to switch off the PC after shutdown and also enables a modem or a LAN card to switch the PC on when it is sent a wake-up frame by a remotely connected machine such as a server or a fax from the telephone exchange. To allow this, the Wake-On-Lan feature needs to be enabled in the BIOS and the device manager.

Checklist before buying a cabinet for your PC

Cabinets today are manufactured with a great deal of imagination. So, there might be a great deal of difference between their shape, arrangement of side panels, etc. The same goes for the motherboard. The cabinet and motherboard shown here are just sample ones, and the positioning of parts and their working will differ across models.

Portable Flash Drives and Memory Cards

USB drives and memory cards are portable data storage devices that use flash memory technology. Flash technology uses a solid-state storage medium and use no moving parts. Flash devices are also smaller and have less power requirements. In USB flash drives have an USB male port to connect to computers. They can be easily plugged in to a USB female port of a computer. You do not have to open the cabinet to install the device. Most modern operating systems treat USB flash drives as removable drives. Memory cards such as SD, SDHC or MMC are meant to be used with cameras and phones. To use them on a PC, the PC needs to have a suitable card reader. Most laptops have a multi-card reader. Desktops do not have this option built-in. However, you can buy a internal or external card reader. An internal card reader is installed on a bay meant for a CD/DVD drive. An external card reader uses a USB port to connect to the computer. Both types have different ports to accommodate different types of memory cards.