Computer Hardware
By Pilot




Computers are an amazing thing.  They come in all shapes and sizes, yet all do the same basic thing.  Through the years, computers have remained the same, stupid machines that they were 30 years ago.  However, unlike 30 years ago, the machines of today are much, much faster.  What used to take an entire floor room to do basic math calculations can now be accomplished with the help of calculators no bigger than the size of your stack of credit cards.  What was once achieved with vacuum tubes is now done with transistors.  What is now being done with electricity will one day be done with light, or other means we have yet to even imagine.

The most popular type of computer, or what we call "the" computer, is the desktop.  It is also called the PC, short for Personal Computer, a term that was used when the first home computers were introduced in the late 1970's and early 1980's.  The desktop computer commonly comprises a monitor, keyboard, mouse, speakers, and a tower - or case.  It is the cheapest form of general computing available, and the most flexible.  With a desktop, one can do almost anything.  From gaming, to programming, to video editing, to song writing, to music editing, to disk jockeying, and so on.  The desktop is also flexible in that upgrading its technology is extremely easy.  Since computers have started appearing in homes, there have only been 2 dramatic changes to the desktop made in the last 20 years or so.  The first was the move to the AT formfactor.  This was the first move to a standard method of desktop computer design.  But at the same time, the AT spec kept some of the legacy features such as you could use the same power supply you already had, for the most part.  Unfortunately, you did have to change case for a model that would support the AT formfactor.  This revolutionized the computer industry.  Gone were the days where each computer maker had their own shape for the case, making upgrading very difficult.  The next move occurred not too long ago, with the introduction of the ATX formfactor.  Once again, this meant moving to a different case for the computer to reside in, but it also brought many advances, such as PS/2 keyboard, USB, AGP, many more.  Unfortunately, upgrading to the ATX spec meant getting a new power supply as the ATX designs were incompatible with the 2-piece connectors used in earlier AT formfactor motherboards.  Of course, transitions such as these do not occur overnight.  And while the ATX spec was being phased in, the AT model was still the most common.  With the announcement of the new ATX formfactor, and in an effort to ease upgrading, AT model motherboards started appearing with the ATX power connector as well as the older 2-piece power connector that was used since even before AT was born.  For the transition from old computers to AT, and from AT to ATX, most of the hardware was still compatible with the newer technologies.  As mentioned, you could still use your old power supply on the newer AT motherboards.  You could still use your expansion cards, your hard drives for the most part, your mouse and keyboard, and to a certain extent, your monitor.  With the move from AT to ATX, once again expansion cards were still good to use, you could still connect your older mouse serial mouse, you could still use your AT keyboard with the help of an adapter.  Hard drives remained usable for the most part, and same goes for monitors.

So as you can see, a lot of hardware in today's desktop computers was compatible with the older AT spec and beyond.  Even now, the desktop is undergoing a slow and uncertain change in formfactor once again.  BTX is the latest formfactor, which you may have heard of.  With the move to BTX however, you will once again require a new case, and possibly a new power supply, we will have to wait and see for the latter.  It is important to note that at the time of writing, it is unsure if the computer industry will accept the move to BTX or not, as the principle reason for the BTX formfactor is to improve cooling to the CPU.

The second most popular type of computer is the laptop.  The laptop is for situations where the use of a desktop would be impractical or impossible.  Such as on the train, in an airplane, or in an environment where you move from one location to the other frequently, or even in conditions which don’t allow for you to haul around a heavy monitor and large case.  The laptop has all the same features as the desktop does, a monitor, keyboard, mouse, speakers, and case.  However, unlike the desktop, the laptop contains all of these inside a small lightweight enclosure.  Laptop computers have the same basic hardware as the desktop computer.

Those two types of computers, the desktop and the laptop, are what we generally call a computer.  However, the word 'computer' means much more.  Your cellular telephone contains a computer.  Most cars contain computers.  Some aeroplanes have computers too.  Your digital watch has a computer and so does your digital alarm clock.  Key word here is 'digital'.  Old analog clocks used metal gears and levers to work whereas digital clocks use computers.  Cars use computers for the electric locks, engine starter, anti-theft device, fuel efficiency for the carburetor, cruise control, and many other components such as ABS brakes.  Big passenger planes use computers to move their ailerons, elevators, and rudders, as well as for their navigation systems, fuel pumps, and black boxes.  Fighter aircraft have computers to assist the pilot in flight as modern fighter aeroplanes are designed with negative stability to make them more agile.  Unfortunately, this means that a pilot would be unable to fly without the aid of a computer.  A computer is also used to display their HUDs (Heads Up Display) and to allow missiles and bombs to lock on to enemy targets.  Computers are even used in your wireless home telephone to send and receive the signals at different frequencies and to display a warning when the battery becomes weak.

The realm of computers is a vast one.  Spanning many different everyday things.  Writing a paper on all computer hardware could take one person their whole life.  This paper is aimed at 'computer' hardware in the sense of desktops and laptops.  While the term 'computer' can be used to describe all that I have mentioned in this introduction, I will stick to using the word computer to describe the desktop and laptop.  Seeing as this is how 90% of the world describes it already.




Table of Contents

1 - The CPU
2 - The motherboard
3 - Memory
4 - The hard drive
5 - Other forms of data storage
6 - Optical drives
7 - Video cards
8 - Sound cards
9 - Expansion cards
10 - The power supply unit
11 - Cooling Solutions
12 - Assembling a computer
13 - Conclusion




The Central Processing Unit (CPU)

The CPU is the brain of the computer.  It is where all the data is processed.  Central Processing Units, also known as just 'processors' have been around for most of the existence of computers.  First generations, such as the ENIAC didn’t have a processor.  The reason behind this is because they were processors.  All the ENIAC did was process data that it received via human intervention.  Technicians of the day had to program the ENIAC before it could process data.  It did not display results via a monitor, nor save them onto a hard drive.  Instead, ENIAC would display results via a series of lights, in the form of binary data.  Once the ENIAC was turned off, all data would be lost.


The first real CPUs were developed by Intel for use in Texas Instrument calculators, in this particular case, it was called the 4004.  While the principle behind CPUs has been around since the very first computer was born, the chip designed by Intel for Texas Instruments was the first time that all the components used for processing data were grouped onto the same chip, which is where the term CPU comes from.  Intel then saw the advantage of using such chips as processors in computers.  Thus, were born the first 8080s and 8086s.  They were followed by the more common 80286, 80386, and 80486, otherwise known by their short forms, 286, 386, and 486.  Each one better than the next.  This was followed by the release of the Pentium line of processors, which still exists today.  The Pentium evolved through the Pentium 1, the Pentium Pro, the Pentium 2, Pentium 3, and now, the Pentium 4.  At around the same time as the development of the Pentium, another company started manufacturing "clones" of the Intel Pentium processors.  This company was called Advanced Micro Devices, or AMD.  Their original chips, called the K5 and K6 processors were simply clones of the Intel technology.  Sold at cheaper prices, and without the Intel name and warranty.  AMD has since gained a large portion of market shares and is now neck to neck with Intel.  With every new line of processors that Intel released, AMD came out with it's own equivalent.  For example, when Intel came out with the MMX instruction set, AMD released the 3D Now! instruction set.  Both accomplish relatively the same thing, but are from 2 different companies.  When the Pentium 3 was introduced by Intel, AMD released the original Athlon.  Both the Pentium 3 and Athlon were short lived, replaced by the Pentium 4 and Athlon XP respectively.  The Athlon XP is now called the Sempron, as AMD reserves the name Athlon for it's higher end class of processors.  As well, just as Intel released low-end versions of it's processors called Celeron, AMD follow suit and called it's low end processors Duron.  Bellow you can see a brief collection of the different processors manufacturered by each company throughout the years, starting with the popular 286, 386, and 486 processors from Intel, and the “clone” released by AMD to the now Athlon 64 which leads the market while Intel is trying to keep up.


Intel and AMD are the two leading processor manufacturers in the western world, Hitachi and Cyrix having been left behind in the dust.  I say western because I’m not convinced that in Russia, you will find Intel and AMD processors in their computers.  But I am not sure.  I don’t live there.  Both Intel and AMD processors accomplish the same functions, and are built the same way.  So now lets take a closer look at the Central Processing Unit.

Processors works by receiving information, processing it, and then sending it back.  Think of a painter.  He receives the canvas and the paint, then turns them into a beautiful masterpiece, and then sells it back.  The processor is given information which it then processes.  It executes various mathematical equations on the data.  Once it is done, it sends the information back, and is ready to receive more information.

All the processor is, behind all that magic that it does, it is millions of switches.  Just like light switches that are either on or off.  That is what the processor is.  A collection of on and off switches.  However, it would be impossible to fit millions of light switches inside the processor die, so instead of everyday switches that you can see and touch, a processor is made of microscopic switches, called transistors.  Like light switches, the switches inside a processor allow electricity to pass through it or not, depending on the state of the switch, on or off.  The pioneering computers utilized vacuum tubes as switches.  These vacuum tubes generated an enormous amount of heat and consumed huge amounts of electricity.  As well, vacuum tubes were big and bulky.  Even Intel's first processors did not use vacuum tubes.  Only the first computers did.  The ENIAC, UNIVAC, Mark 1, etc.  But these computers were floor space giants costing millions of dollars simply to accomplish basic math.  The transistor later replaced the vacuum tube due to its smaller size, low power consumption, and almost no heat generation.  Transistors are the switches that are in use today in processors.  Simple microscopic on/off switches.  When Intel released its 8080 line of processors, it contained 4 500 transistors on a chip the size of your finger nail, and operated at a speed of 2Mhz.  Compare this to the Pentium 4 processors of today which average around 125 million transistors on a die the same size as the 8080 and operating at speeds of over 3Ghz!

However, unlike the early ENIACs and UNIVACs, processors of today are divided into sections, and each section has specific functions.  For each section to perform it's specific duty, it must be made differently from other parts of the processor, yet still using the basics of transistors.  The ENIAC didn’t have the luxury of a floating point unit, or an L2 cache.  All the vacuum tubes were all of the same fashion, and with hundreds of them,  ENIAC was able to compute addition, subtraction, division, multiplication, and so on.

Processors of today are divided into sections, which compute different forms of data to be more efficient.  Data is under the form of binary code.  The only language the computer can truly understand.  Binary data is comprised of 1s and 0s.  This is because of how computers are made.  With switches that can either be on or off.  Therefore, 1 or 0.  However, with multiple switches, you can make all the numbers imaginable.  And it is this way that the computer operates.  With different numbers.  And with those numbers, it can make other numbers which are then equivalent to letters and symbols, or to pixel colours which appear on your monitor.  Each 1 or 0 is called a bit.  So a transistor which allows electricity to pass represents a 1 bit of data, and a transistor which is closed, represents a 0 bit of data.

Data on most of today's CPUs are processed 32 bits at a time.  This is what we call 32-bit processing, which is found in almost all processors today, with the exception of AMD's Athlon64 and FX series of processors, and some of Intel's Xeon processors.  What this means is that the processor can work with 32 bits of data every clock cycle.  Meaning 32 transistors, 32 1s and 0s.  While 32 bits of information isn’t a lot, it is more than made up for with clock speed.  As I mentioned, 32 bits are processed every clock cycle.  But there are millions of clock cycles each second.  This is represented by the term Hertz.  1 Hertz is 1 cycle per second.  1000 Hertz is equivalent to 1 Kilohertz (Khz), and if you continue the pattern, 1000 Kilohertz is equal to 1 Megahertz (Mhz) to which 1000 corresponds to 1 Gigahertz (Ghz).  So a processor operating at 1Mhz executes 1 million cycles per second, resulting in 32 million bits of data processed per second.  However, today's processors do not operate at 1Mhz.  They operate between 1Ghz and 3.8Ghz, which makes for a lot of data processed each second.  This is why computers are "incredibly stupid, yet incredibly fast".  Because you won’t believe how many operations the processor can do each second, and you wont believe me when I tell you all it understands is either a 1 or a 0.

Processors have two speeds.  An internal one and an external one.  The internal one is the speed at which the CPU performs operations on data.  It is the most common speed that we refer to.  The external speed is how fast the processor can communicate with the motherboard - and the rest of the computer - and is called the Front Side Bus, or FSB.  A Pentium 4 3.8Ghz with a 800Mhz FSB has an internal speed of 3.8Ghz - or 3800Mhz - and an external speed of 800Mhz.  The external speed is determined by the motherboard, from which the processor is constructed to work at that speed.  The internal speed is determined by the motherboard as well and is always a multiple of the external speed.  The processors are pre-programmed to run at a specific multiplier which is why a 3.8Ghz Pentium 4 its speed of 3800Mhz.  Take for example the Athlon XP 2500+ with a 333Mhz FSB and a speed of 1.833Ghz.  Since the 333Mhz speed is determined by the motherboard, the variable speed is the internal speed of 1.833Ghz.  To achieve an internal speed of 1.833Ghz, the FSB must be multiplied by a factor of 5.5 if we round to the first decimal number.  If you multiply these, you will get 1831.5 which would then be the internal speed of the processor.  The options for multiplier and FSB can be configured, to a certain extent, via the BIOS menu, which can be loaded up when the computer is in it's Power-On Self Test (POST).  Inside BIOS, you can modify the speed of the Front Side Bus, thus overclocking the CPU, motherboard, and RAM - unless the RAM is capable of running at a higher FSB already.  Increasing the FSB can be dangerous and can lead to premature failure of any of the three mentioned components.  Overclocking also voids your warranty and requires additional cooling.  When you increase the FSB, you are increasing the external speed of the processor (which is also called the FSB), and since the internal speed is a multiplier of the external speed, you are indirectly increasing the processor's speed.  This is the easiest way to overclock a system, and, if you don't know what your doing, the best way to destroy your system too.  The second way you can overclock a processor is to increase the multiplier.  This will increase the CPU's internal speed, but will not affect the FSB.  This method can yield higher processor speeds, but can cause some damage to the CPU if it is not cooled proprely.  Increasing the multiplier has no direct effect on the motherboard or RAM.  I say direct because if you burn your processor, odds are you will damage the motherboard and most likely the system memory as well.  Increasing the multiplier is an easy way to gain around 20% more speed from a processor, and this resulted in some people purchasing lower-rated CPUs and increasing their multipliers and then reselling them as higher clocked processors.  Doing this several hundred times would gain them considerable amounts of profit.  Intel and AMD were both hurt by this form of piracy and locked multiplier speeds on most of their CPUs.  Most Pentium 4's and almost all Athlon XP's have their multipliers locked.  There are also many other ways of overclocking a CPU, but they will be discussed later in this section.




The Motherboard

The motherboard, also known as mainboard, is where every piece of hardware is directly or indirectly connected.  Think of the motherboard as a city grid, with streets and intersections.  Motherboards, have always existed  in computers, with the exception of pioneering computers such as the ENIAC and Mark 1, which were simply just giant processors.  The motherboard is always in a state of evolution as new technology comes out.  The reason behind this is because the motherboard is the backbone of your computer.  It is the only component that you need technically, to have a computer.  Mind you, you wont get very far without a processor, monitor, hard drive, and ram, but you can still accomplish something with only a motherboard and a power supply.  You cant accomplish anything with just a hard drive and power supply.

Through the years, the desktop motherboard has undergone countless changes.  Some more significant than others.  Such as the transition to the AT (Advanced Technology) formfactor.  Formfactor meaning the shape, and the location of key components, such as the processor socket, I/O ports (Input/Output), expansion slots, and system memory slots.  AT was the first real formfactor for motherboards.  The AT formfactor placed the processor socket in the same area to allow easy access to it, and to ensure that processors and their heatsink and fan combos would be able to fit in any case with the AT spec, thus doing away with the worry of incompatibilities due to physical limitations.  I/O ports were all found in the same location so that all computers would be plugged the same way.  As the years went on, the AT formfactor started to show it's age.  With the introduction of the PS/2 keyboard and mouse, and with the introduction of USB and many other new technologies.  This pushed the introduction of a new formfactor called ATX (Advanced Technology Extended) which again, forced the desktop industry to change.  The old AT cases and power supplies were physically incompatible with the ATX spec motherboards.  ATX introduced the PS/2 keyboard and mouse ports, as well as integrated serial and parallel ports instead of the ribbon cables that were needed with the AT spec to connect them to available expansion slots.  ATX also introduced the USB (Universal Serial Bus) ports, and had IDE (Integrated Drive Electronics) ports built in to all models.  Previous AT motherboards sometimes required PCI (Peripheral Component Interconnect) or ISA (Industry Standard Architecture) controller cards to plug your hard drives to.  The ATX formfactor also moved the location of the processor socket.  With the coming of more powerful CPUs, the need for additional cooling was becoming greater.  While you could run a 486 chip without even a heatsink, new Pentium processors required both heatsink and fan.  Newer Pentium 2 processors generated more heat, and it was assumed that as processors evolved, the amount of heat they generated would be greater.  Therefore, ATX placed the processor socket in the upper part of the motherboard where it could receive additional cooling from the power supply's fan.  It also placed the socket away from the drive bays to allow more physical space for larger heatsinks on newer processors as well as allowing the large Pentium 2 processor cartridge more room.  While AT standardized the location of the I/O ports, expansion slots and general shape of the motherboard, ATX did not change this design.  It simply improved on it to allow for newer technology, and planning ahead for future technology.  Now that the ATX spec is almost 10 years old, some people are starting to think that it is becoming impractical.  Brand new is the BTX (Balanced Technology Extended) formfactor, which has been talked about for almost a year now.  This new formfactor would once again mean your case is incompatible with the design of the new motherboard.  However the reason for BTX - and the reason why its not being accepted to quickly in the industry - is that it was only created to aid in the cooling of the processor.  BTX formfactor is the same as the current ATX spec, but it is inversed.  Take a mirror to your ATX motherboard, and you will find yourself with the BTX formfactor.  The only difference in layout, is that the processor socket was moved from the top of the motherboard to the front of the motherboard.  With newer cases featuring as many as 5 to 7 case fans, the CPU would gain more cooling by being placed in front of 1 or more intake fans than it would being in the airflow of the power supply.  This is because when the ATX spec first came out, cases didn’t come with intake or exhaust fans.  The only two fans in the system were on the power supply fan and the processor fan.  Nowadays it isn’t uncommon to find cases with 7 to 10 fans in them.  My own case has 9 fans.  Two in the power supply, one on the processor, one on the northbridge, and five case fans, three intakes, and two exhausts.  My video card had one but this has been replaced with a Zalman heatpipe/heatsink and the northbridge fan will be replaced with a Zalman heatsink as well.  With so much heat generation, and with the processor socket on the ATX motherboards being placed at the top and the back of the motherboard, the air that was flowing over it would be already warmed up from being passed over other components such as the hard drives, video card, northbridge, and RAM.  The BTX formfactor moves the processor socket to the front of the motherboard at mid level, right where most case intake fans are located.  This means that the processor receives the cool air first, before any other component, other than the hard drives.  The BTX spec also moves the motherboard from the right side of the case to the left side, meaning that to work on the computer, you would need to remove the right side panel instead of the left side, like with the current ATX spec.  It is also unknown if current ATX power supplies will be compatible with the BTX design.  This, with the need for a new case, causes more worry in changing to the new formfactor, seeing that the only advantage is to allow better cooling to heat-critical components such as the processor.  BTX does not offer advantages to systems using water or phase change cooling.  This is why BTX has been around for almost a year, but not released for general usage.  More on system cooling in a later section.

It is important to mention at this point, that AT, ATX, and BTX are not the only form factors.  Even with the introduction of AT that standardized the shape of the motherboard, some companies such as Dell and Compaq were still quite happy producing their own motherboards to fit in their cases.  They continued doing this until the early life of the ATX formfactor, when it was finally found to be too much trouble and harder to keep up with companies which designed better motherboards.  Dell and Compaq, amongst others, often had small cases for which the monitor would come to rest on, as can be seen in the picture below.  The motherboard would lie flat on the floor of the case, however the case was not tall enough to accept the full hight of a PCI or ISA card.  Therefore another card had to be inserted into the motherboard to which expansion cards would be connected to.  This form of motherboard design is called NLX, while it doesnt stand for anything in particular, NLX would be more closely abbreviated for New Low profile Extended formfactor.  The second card which connects the motherboard and the expansion slots is called the riser card, and is always mounted perpendicular to the motherboard, which means expansion cards are parralel to the motherboard.

As you can see, over a period of many years, the desktop motherboard has only undergone 2 drastic changes, making it necessary for a full system upgrade.  The move to AT, and the move to ATX.  Laptop motherboards are a lot more unforgiving.  Unlike desktops, laptops are still not fully standardized, other than processor, hard drive, and RAM.  Laptop motherboards still come in all shapes and sizes, and are incompatible from one laptop to the next, unless you purchase a laptop of the same make and model.  Even then, removing a motherboard from the laptop is near impossible unless you know full well what you are doing before hand.

So now that we know the history of the motherboard, let us take a look at some of the common components that can be found on them.  Motherboards are divided into 2 sections, the top half, and the bottom half, commonly called north and south.  In the northern half, you will find the processor socket, RAM slots, AGP (Accelerated Graphics Port) slot, northbridge, I/O ports, and power connector(s).  In the bottom half, the southern part, is located the PCI expansion slots - and if you own an older ATX motherboard, ISA slots - , BIOS (Basic Input/Output System) components, southbridge, IDE connectors, front I/O headers, and USB and firewire headers.  Each half of the motherboard is controlled by a chip, either the northbridge and the southbridge, which takes care of functions in their respective halves of the motherboard.  The only exception to this is the I/O ports that, while located in the northern half, are controlled by the southbridge.  This is made so that high priority devices are located on the northern half.

The northbridge and southbridge act as intersections connecting all the major components mentioned above together, with the exception of the BIOS components.  When you hear the word 'chipset' it is most commonly used to refer to the northbridge and southbridge chips on the motherboard.  The most common one today is the nForce series of chipset made by nVidia.  These chipsets are found on AMD processor motherboards only.  Other chipset makers include VIA and Intel, and a few other companies.  Newer AMD 64-bit processors do away with the common northbridge and southbridge however, and only use on chip for the whole motherboard.  But since most people do not use 64-bit processors, I will not discuss about their chipset, and will instead, stick with north and south.

The northbridge is now found with either a heatsink on it or a heatsink and fan combo, and is located just bellow the processor socket.  It ties together the processor, RAM, AGP or PCI-E 16x slot, and southbridge.  In modern computers, all 3 of those components are high priority devices which require the most speed out of the system, and need the fastest connections to the CPU.  Due to the amount of speed required, and the amount of work the northbridge must do, it is covered with a heatsink and sometimes a fan to cool it down.  The southbridge handles to rest of the principle components which don’t have high speed and priority requirements, such as the hard and optical drives, PCI and ISA slots, I/O ports, and integrated sound if it is featured on the motherboard.  Due to the lighter load of work the southbridge must do, it does not have a heatsink to cool it down.

[INSERT DIAGRAM]

As we discussed earlier, processors have 2 speeds, an internal speed and an external speed.  Motherboards have a FSB speed and so do RAM sticks.  All three components must be able to run at the same speed.  This means that if your motherboard has a FSB of 333Mhz, then the CPU and RAM must be able to run at 333Mhz as well.  The FSB speed on the motherboard is the speed at which the northbridge operates.  Since, as shown in the above diagram, for the processor to communicate with the RAM, it must go through the northbridge.  The reason for this is because the northbridge contains the memory controller.

Newer 64-bit AMD processors have the memory controller integrated into the CPU itself.  This is why the north and south bridge chips are not used.  It allows the FSB speed of the processor to be equivalent to the internal speed of the processor.  Therefore, a 64-bit processor running at 2.6Ghz would have a FSB of 2.6Ghz since communication between CPU and RAM is done directly, without having to go through a slower chip on the motherboard.  The only disadvantage to this is that current memory is not fast enough to keep up.

Since the CPU is fed data from the RAM, it is important that they have as fast as possible a connection together, which is why they are connected together via the northbridge.  For games and graphic intensive applications, the video card comes into play.  The GPU (Graphical Processing Unit), also known as the VPU (Visual Processing Unit) on the video card performs billions of calculations per second from data and instructions that are stored in it's RAM.  It gets this information from the CPU, so yet again, it is important for both of these to have a fast connection to one another.  That is why the video cards is tied to the northbridge as well.  There are many other functions on the motherboard which aren’t as high priority as the processor, RAM, and video card.  These are tied together to the southbridge, which is then linked to the northbridge.

The hard drive is much much slower than the speed of the RAM and processor, so, even though it is a device that is in constant usage, it would just be impractical having it connected to the northbridge.  At the same time, although you, as a user, might be typing on the keyboard constantly and using the mouse, in the time frame of the computer, it isn’t a commonly used device.  That is why all I/O ports are on the southbridge.  This is because we live our lives in seconds and minutes, while the computer lives in nanoseconds.

Items such as the keyboard, mouse, hard and optical drives and expansion slots are all connected to the southbridge due to lower priority.  Your keyboard doesn't need urgent use of the CPU, and can wait a few clock cycles without the user noticing any form of lag.  Same goes with the hard drive.  As fast as a 15 000rpm drive can be, it is still way too slow to need priority CPU usage.  Since all these devices are lower priority, they have been tied together, seperate of the RAM, CPU, and high end video cards.  The southbridge is an intersection between all of those devices and links them to the northbridge.  On motherboards that have integrated sound, such as the nForce2, the sound card is integrated into the southbridge.  This adds additional tasks to the southbridge, even more so if the integrated is an accelerator, which means it does not offload processing to the CPU.  This means the southbridge is also an Audio Processing Unit, or APU.  The same goes for integrated video, however this is a task held by the northbridge.  Integrated video cards are always accelerators bassed on a current model of video card, such as the popular GeForce4 MX series of graphics.  Integrated video is accomplished by combining the northbridge with a GPU, which will then use system memory as it's onboard RAM.  This gives the integrated GPU extremly fast access to the CPU to receive instructions, just like a standard video card would, however, as fast as the system RAM is, it is still slower than the onboard memory of today's video cards, which means overall, integrated video cards are slower.




Memory

What is memory, you may ask.  Well, it is the ability to remember something, to store information.  Whether temporarily or permanently.  Inside a computer, there are many different forms of memory.  And it is found almost everywhere.  The most common type of memory we refer to however, is system memory.  But we most commonly call it RAM.  However, it is important to note that the real name for it is in fact 'system memory' whereas RAM means the type of memory.  RAM stands for Random Access Memory.  What this means, is that you can access any part of data on the memory without having to go through everything else first.  The tape drives of yesterday that were used for data storage were in fact a type of memory, but they were not RAM.  They are called SAM for Serial Access Memory.  Because to get to a certain point, you had to go through everything that came first.  Just like you have to fast forward through a VHS tape or a music tape, all of these, SAM type memory.  Those being phased out my RAM-type devices, the CD and the DVD.

Another common type of memory is ROM memory.  Read-Only Memory is most commonly found in CDs, hence, CD-ROM.  What this means is just as the acronym suggests.  The data contained on the media is in a permanent 'read-only' state.  You cannot alter the data.  CDs were originally as such because the data couldn’t be altered after it was stamped at the factory.  Same goes with DVD-ROM media.  However, ROM also finds itself in the solid-state world.  For example, the BIOS chip on your motherboard is a ROM module.  The data on it cannot be altered... to a certain extent.  But more on that later.

The opposite of ROM is also true.  Writable memory is only found in optical media, and denotes a blank disk to which you can only write data to it once.  Writable media is labeled with the suffix R, as in CD-R.  Rewritable memory, or RW, is once again only used on optical media, such as CDs and DVDs to mark the difference between a ROM disk and a RW disk.  Conventional system memory is, of course, RW, since you can rewrite to it at any time.  So it would be impractical to add the RW suffix to it.  Optical media, however, can be either one.

Those above suffixes are the different forms which memory can be found.  RAM, ROM, R, and RW.  But there are many different types in which memory can be.  The ones we refer to as RAM, or system memory, are integrated circuits (IC), or chips.  Other forms of memory include optical media, floppy disks, and even your hard drive.  But each of these will be covered in their own section.

The form of memory that will be covered in this section is the system memory, commonly referred to as just 'RAM'.  A computer's main memory, or system memory, or RAM (depending on your preference) is one of the most vital components in the computer.  Everything you do and see on the monitor comes from the RAM.  So without RAM, you couldn’t do much.  Information enters your RAM, and then is processed by the processor before being sent off to wherever it has to go, such as to the video card, to be displayed on your monitor.

System memory comes in different flavours, just like ice cream.  RAM can either be Static or Dynamic.  So either SRAM or DRAM.  Static RAM means that it will keep it's settings until they are changed again, regardless of power state.  This means that the computer could be turned off, and the data in the SRAM would still be intact.  DRAM needs to be constantly refreshed to keep its data.  So if you turn off your computer, you loose all the information in the DRAM.  All computers use DRAM as main memory since it is cheaper than SRAM.  SRAM, however, can be found on your BIOS chip, your MP3 player, your Nintendo game cartridges, your USB key, and many other places.  Since DRAM is much cheaper and more readily available, it has undergone constant evolution to keep up with higher demands and performance requirements.

RAM stores data, just like every other component, in binary.  Meaning 1s and 0s.  This is accomplished via different methods depending on the type of RAM, either SRAM or DRAM.  SRAM uses a series of on/off switches, like light switches to store information.  The same way that processors use switches.  Electricity is used to move the switch one way, to represent a 1, or the other way, to represent a 0.  And the data is kept until electricity is used again to change the binary number.  DRAM uses very small capacitors instead of switches.  If the capacitor is holding a charge, it represents a 1, if it is not, it represents a 0.  This is why constant electricity is required, so that the capacitors can continue holding their charge.  SRAM requires four different switches to store 1 bit of data, while DRAM requires only one capacitor to store a bit.  This makes DRAM much more practical than SRAM and the better choice for system memory, as memory cells can be much smaller.

DRAM has been around for many many years, and has gone through many different changes.  Without going into too much detail, the first types of DRAM, which are now completely obsolete.  The first type of DRAM used in desktop computers was called FPM RAM, for Fast Page Mode RAM, and was first introduced in 30-pin SIMM sticks and carried through to the 72-pin SIMMs.  The next evolutionary step was EDO RAM, which stands for Extended Data Out RAM which started and remained on the 72-pin packaging, and carried through both the 72-pin SIMM and 72-pin DIMM.  The reason I say this is because, at the time, DIMM was an improvement to SIMM, since the technology didn't yet exist to make better IC chips that could contain larger amounts of memory cells.  Typical ones today can hold 32mb on each chip.

The next step is the old - yet still very used - SDRAM.  SDRAM stands for Synchronous Dynamic Random Access Memory.  It was the first type of DRAM to run at the same speed as the computer.  It was smart in that it would synchronize itself with the speed of the motherboard and processor, the Front Side Bus.  This means that the flow of data between the memory and the processor is constant.  Older types of DRAM were either too fast, meaning that they had to wait and hang while the motherboard sent the data to the CPU for processing, or they were too slow, which meant the processor was wasting valuable cycles doing nothing.  SDRAM generally runs at speeds of either 100Mhz or 133Mhz but adapting to the motherboard's speed.  SDRAM which runs at 100Mhz will not be able to run at 133Mhz if the motherboard can do so.  Instead, the motherboard and processor will have to clock down to 100Mhz to match the speed of the RAM.  This is why it is important to have RAM which matches or exceeds the speed of your motherboard.  You can think of clock cycles as wavelengths, just like radio wavelengths.

[INSERT WAVELENGTH]

Each wavelength represents one cycle.  Or one hertz.  While a wavelength has both a upper part and lower part to it, SDRAM only uses one part of the cycle for data transfer between it and the motherboard.  This limited SDRAM to the amount of data it could transfer, and the only way to increase it, was to increase the speed.  But this was impractical, so a new form of DRAM had to be created.  SDRAM started appearing at the same time as the AT formfactor, and to ease the flow of technology, AT motherboards came with both 72-pin slots and 164-pin slots for EDO or SDRAM.  You couldn't use both types of RAM at the same time though.

DDR SDRAM was the next step in the ever so evolving state of DRAM.  Double Data Rate Synchronous Dynamic Random Access Memory was an improvement to the older SDRAM.  DDR SDRAM has 2 major improvements over SDRAM.  The first, is it's ability to go to higher speeds, ranging from 300Mhz to 400Mhz.  The second feature, and most important one, the reason why it is called 'Double Data Rate' is because it sends data on both the rising and falling parts of the clock cycle.  So for each wavelength, data would be sent twice, at the crest and at the trough.

Newer DDR2 SDRAM has appeared only a short time ago, but hasn't really become popular due to its high cost.  DDR2 RAM uses 240-pin SIMMs and DIMMs making them once again incompatible with 184-pin DDR slots.  DDR2 SDRAM clocks between 400Mhz and 667Mhz, improving over the speeds of current DDR.

In the realm of video cards, DRAM has continued to evolve at a more aggressive pace.  Current high end cards utilize either DDR3 SDRAM or GraphicsDDR3 SDRAM, and have much higher clock speeds to give better performance for the billions of calculations the GPU must do every second.  Slowly, DDR3 SDRAM will find itself in system memory when the technology curve catches up, and when prices go down.

The main memory, over the years, has evolved.  And with the newer features and faster speeds, needed newer methods of integrating the RAM with the rest of the computer.  Many years ago, back in the days of the 286 and before, RAM was not what it is today.  While RAM back then did the same functions, it was not a PCB stick with IC chips on it that you would put into expansion slots.  Back then, RAM was put directly onto the motherboard in a fashion called DIP (Dual Inline Package) memory.  This quickly became obsolete as upgrading the RAM meant getting a new motherboard.  But back then, you couldn’t simply get a new motherboard, you would have to get a whole new computer.  Another disadvantage of DIP memory was that it consumed huge amounts of floor space on the motherboard.  Almost as much as your series of PCI slots take up on your motherboards, imagine that full of DIP memory chips for a total of perhaps 2mb of RAM or so.  Not very practical at all.

Next in the evolutionary chain was the introduction of the SIMM.  Single Inline Memory Module was first introduced in 30-pin sticks with RAM modules on one side only, hence the name 'Single'.  This revolutionized the computer industry as it was the first time a computer could really be upgraded.  Several slots on the motherboard permitted the introduction of more SIMMs to add more RAM to a computer, without the need for a new motherboard.  Once RAM became too performing for the 30-pin SIMM, 72-pin SIMM modules started appearing.  They progressed through the AT formfactor as well, but were starting to be phased out at that time by the newer and faster SDRAM.  30-pin SIMMs run at 8bits, meaning that every clock cycle, it would receive or send 8bits of data.  72-pin SIMMs and DIMMs functioned the same way, but at 16bit.  SDRAM was so advanced that using even 72-pin SIMM wasn’t enough.  SDRAM comes in 168-pin SIMMs but only to a certain extent.   The problem at the time was that limited technology couldn't fit large quantities of memory on single IC chips, therefore creating larger quantities of RAM on the same stick became impossible.  The next step was a simple one.  Why not add memory modules to the other side of the stick.  Thus, DIMM was born.  Dual Inline Memory Modules first came out with the EDO RAM on 72-pin and allowed the chips to be mounted on both sides, effectively doubling the amount of RAM that could be contained on a single stick.  However, it is important to note that DIMM did not replace SIMM.  SIMM is still in use today.  Now that we can have ICs with up to 32mb, SIMMs are once again proving useful as now you can easily fit 512mb of system memory on the single side of a 184-pin DDR SDRAM module.  A few years ago, a 168-pin SDRAM module containing 32mb required the use of DIMM as it couldn't be done using SIMM due to lack of technology.  For example, two out of three memory sticks in my computer are SIMMs and the last, DIMM.  Another lesser known form of configuration is Rambus Inline Memory Module.  RIMM is both a packaging and a memory type, introduced a little after SDRAM.  RIMM is extremely fast, at 800Mhz, still faster than DDR2 SDRAM today.  Yet it was never popular.  RIMM came in 184-pin DIMMs and looked like a DDR SDRAM 184-pin but was slightly longer.

As modern technology advanced, and laptops became more common, the type of RAM they use became standardized.  While they still use SDRAM and DDR SDRAM, their method of delivery isn’t the same.  You cant fit even a 30-pin SIMM module in a laptop, and much less a 168-pin SDRAM module or even a 184-pin DDR SDRAM stick.  Laptops use their own SoSIMMs and SoDIMMs slots.  Small Outline Single - or - Dual Inline Memory Modules are like miniature versions of SDRAM and DDR SDRAM SIMM and DIMM modules, made specifically for laptops.  They offer the same features, but in a much smaller package.  And having a smaller package means a higher purchase price.

So now that we know more about RAM, lets take a look at how it does it's magic.  How is it that a capacitor can mean a 1 or 0.  How does the computer access information in RAM at "random", and so on.

DRAM is made up of millions of transistors and capacitors, each grouped together, one of each, to form what is called a memory cell.  Each cell holds 1 bit of information.  The capacitors are unique to DRAM.  SRAM doesn't have capacitors.  The capacitors in DRAM are what actually store the data.  Capacitors work by holding a charge of electricity, but, like a open wound, they slowly bleed.  However, instead of bleeding blood, which is our life liquid, they bleed away the electricity.  This means more electricity has to be put in to compensate for the loss, which is why DRAM needs constant electricity to work.  Because, while capacitors can store a charge, it slowly bleeds off.  Which is why, when you turn off the power on your power supply, that the motherboard light stays on for a while, because the capacitors within the unit are bleeding away their electricity.  This is where the transistors come in.  With the transistor open, electricity is allowed to enter the capacitor, representing a 1bit, and with the transistor closed, electricity is cut, and the capacitor bleeds away, representing a 0bit.  The capacitors in DRAM are so small, that the time needed for the capacitor to bleed away is mere nanoseconds, and not 3 to 4 like those beefy ones in your power supply.  Electricity is only needed once to turn the transistor on or off, but constant electricity is needed for the capacitor.

Each integrated circuit chip is made up just like a grid, with rows and columns.  These are called CAS and RAS, which you might have heard about elsewhere.  Column Address Select and Row Address Select are pipelines going through the IC chip.  If you compare a DRAM integrated circuit chip to a piece of graph paper, the horizontal lines would be the RAS and the verticle lines would be CAS, while the empty squares would be memory cells, meaning a transistor and capacitor.  Each memory cell is given a specific address, depending on which CAS and RAS it is located on.  Here is an example:

[INSERT PICTURE HERE]

The CAS would be connected to the transistors while the RAS would be connected to the capacitors.  The memory controller, usually located on the northbridge, is what controls the entry of data into the system DRAM.  For data to be written, the first thing that happens is that a charge is sent to the CAS, which turns on all the transistors in that column, thus allowing any electricity to pass from the RAS to the capacitor.  In the second step, specific RAS pipelines are charged, which sends electricity down the pipeline, until it reaches the column of memory cells which is currently active due to the charge sent down the CAS pipeline.  The active RAS pipelines which intersect the CAS pipeline then charge the capacitor in the memory cell.  Once that process is done, the next CAS pipeline is charged, turning on all the transistors, and again, certain RAS pipelines are charged, which in turn charges another set of capacitors, resulting in more data being written.

[INSERT PICTURE]

However, since capacitors slowly bleed away their electricity, the capacitors must be refreshed, meaning that the CAS and RAS pipelines for that cell must be re-activated after a period of time to recharge the capacitor in order to maintain it's data.  You can think of it somewhat like the refresh rate on your monitor.  Except RAM cells must be refreshed much more often.

For data to be read, each capacitor is measured for the amount of charge it contains.  If the capacitor is over 50% charged, it is interpreted as a 1bit, and if the charge level is bellow 50%, it is considered a 0bit.  The CAS will charge a column of transistors.  While the process seems complex, it is done incredibly fast.  During the life of DRAM, when the computer is turned on, data is written to the memory cells, read, and re-written.  That is how the capacitors maintain their charge.  This is what is called the refresh rate of the memory.  The speed at which it refreshes is expressed in nanoseconds, so you can get an idea of how fast memory is.  Each time the memory refreshes, if data in the memory cells need to be changed, this is when it can be done.

The CAS and RAS pipelines, as well as the memory cells, are all imprinted on a silicon waffer, just like the CPU.  These small silicon chips, are then placed inside protective plastic casings, which we call IC chips, for integrated circuit.  IC chips are of predefined sizes, which can be read off the sticker on the memory stick.  For example, in my laptop, I have a SoDIMM module for 128mb SDRAM.  Each side contains 4 ICs, meaning there is a total of 8 integrated circuits, which provide the 128mb total memory.  On the sticker on the side of the RAM stick, you can see how the memory is organized, in this case, my SoDIMM module says "16Mx64".  This means that each IC chip contains 16mb of memory arranged along 64 Column Address Select pipelines, which, physically go up and down the longer part of the IC chip.  From there you can figure out how many RAS pipelines there are by calculating the ammount of bits in 16mb and then deviding that number by the number of CAS pipelines, in this case, 64.  There are a total of 134 217 728bits in 16mb, therefore deviding 134 217 728 by 16 gives 8 388 608 which would mean that each IC chip on your RAM module contains 64 CAS pipelines and 8 388 608 RAS pipelines.  This means that each chip contains 134 217 728 memory cells, to which each hold one bit of data.

SRAM is a little different.  For one, they do not have tiny capacitors used to store information.  SRAM uses transistors to store the information.  DRAM has one transistor and one capacitor per cell, while SRAM has four transistors per cell.  This makes SRAM cells much larger than DRAM.  In the long run, this means that less data can be stored in the same sized IC chips.  However, since SRAM does not use capacitors, it doesn't have to spend time recharging them, making SRAM much faster than DRAM.




The Hard Drive

We mentioned earlier that the hard drive was in fact, a form of memory.  Well, it is.  But the hard drive is a different form of memory.  It serves a different purpose.  Which is why it has its own section.  Because the hard drive is truly unique, and far different from any other type of computer memory.

Hard drives have been around for a long time, but not as long as the first computers.  The ENIAC and Mark 1 never had a hard drive.  The PDP series of computers never had hard drives either; They used large tape drives.  When the ENIAC computed a series of equations, it would display the results afterwards.  At this point, all processing would stop and this is what allowed humans to view the results.  When the next computation was entered, the original results would be erased.  During the time of ENIAC, this wasn’t a problem.  But as the years went by, the need for some form of data storage became apparent.  In come the massive tape drives and original 5.25inch floppy drives, also known as 'B' drives.  Later, the first hard drives were born.  Their principle was based on that of the floppy drive, meaning a disk rotating onto which data is written in the form of 1s and 0s.  Hard drives received their name in the fact that their disks, known as platters, were hard compared the floppy disks which were flexible, hence the name floppy.  The first hard drives were monstrosities, weighing tonnes, and boasting several platters, sometimes more than a dozen, only to store mere megabytes of data.  Compare this to the laptop drives of today which measure about a half inch tall using two platters and containing 80 gigabytes of information.

Hard drives are, however, one of the simplest devices found in the computer.  Trying to understand how Integrated Circuit (IC) chips - such as the CPU and RAM - work is much harder, yet 80% of components in the computer are IC chips.  The hard drive is also the only component that has undergone very little change over the years.  In fact, the most drastic change to the hard drive over the years has been the decrease in the drive's physical size.  Hard drives found in 10 year old 486 machines can still work on your modern Athlon FX-55 server.  Although the 'time since last defrag' might be in the thousands of days which could make someone worry until you tell them the hard drive is from your old 486.  They are the most compatible components in the entire computer.  They survived the transition to the AT formfactor, and to the ATX spec, and, if ever BTX becomes mainstream, it will survive that as well.  Meaning you could find your old 486 hard drive in a BTX motherboard computer 5 years from now.  We are talking about technology that could be over 20 years old by then, and still kicking!

The inner workings of a hard drive are very simple.  A spindle motor rotates one or more platters at high speeds while an arm moves over the platter to read or write the data.  The first ever hard drive operated that way, and the bleeding edge 15 000rpm Seagate server drives operate the same way.  Let us first tackle the spindle.  A small high powered motor is enclosed inside a shaft, and is attached to the base of the hard drive's case.  This motor spins at a constant speed, thus rotating the platters at constant revolutions.  Early drives spun at 4 300rpm, then went up to 5 700rpm, and now reside at 7 200rpm.  The faster the spindle, the faster the platters can move under the read/write heads.  Newer Raptor drives from Western Digital rotate at 10 000rpm and some, such as the new Seagate server drive, rotate at 15 000rpm.  Increasing the revolutions per minute is the best way of improving the performance of a hard drive.  However, despite these increases in speed, the hard drive still remains the slowest component in the computer by a wide margin.  Modern hard drives utilize between one and four platters which are mounted on the spindle at equal distances from each other.  Platters are simple disks only 2 or 3 millimeters thick made of metal or glass to which the surface is covered in magnesium particles, and then covered with some form of protective coating.

Hovering mere microns away are the drive's read and write head.  At the end of a long arm, the head is a simple metal horseshoe to which a wire is weaved around it.  There is an arm and head for each side of the platter.  A drive containing one platter has 2 arms and 2 heads, while a drive containing three platters has 6 arms and 6 heads.  No matter how many arms a hard drive has, they are all connected to the same shaft which rotates along the same pivot point.  On the other end of the arm is a type of counterbalance made of metal.  It is surrounded by a large magnet, called a voice coil.  This magnet is made of a rare type of metal and is extremely powerful.  It is this magnet which moves the drive's arm over the surface of the platters.  The magnet receives different electrical currents which move, at lighting speed and with deadly precision, the read and write heads.  While the arms are moving, you can often hear the noise, which sounds like a very rapid clicking.  Each click is when the head moves and arrives at it's destination.  When the head is in use, either moving, reading, and writing data, it is called 'disk activity' and it causes the red light at the front of the case to light up, indicating the drive is in use.

Hard drives can be described, very faintly, as RAM.  They are in fact, Random Access Memory devices.  However the term RAM is reserved for solid state memory.  The reason for this comparison is that hard drives can access any part of their data at random, without having to go through the first part of the data.

Hard drives store their data in tracks, which are concentric circles that go along the platters.  Typical hard drives contain thousands of tracks on platters 3.5inches in diameter.  From there, each track is divided into sectors.  Sectors usually hold 512bytes of information on them.  The first hard drives cut sectors into tracks just like you would cut a pie.  This meant that sectors on the outer tracks were larger than the sectors on the inner tracks, but still holding the same amount of data.  This was a horrible waste of space.  To fix this, a method called 'zoned recording' was conceived where tracks on the inner most track and outer most track were the same physical size.  This meant that the outer tracks held more sectors than the inner tracks did.

The drive's read/write head moves over sectors of the disk and either read or write data to it.  For a hard drive to write data to the platter, an electric signal is sent over a wire to the drive's head, which is a piece of metal in the shape of a horseshoe, with a metal wire spun around it to form a coil.  If the electric signal is sent one way, the metal horseshoe becomes a magnet and will orient magnesium particles on the platter one way.  If the electric signal is sent in the opposite way, the head magnetizes the other way, and the particles are oriented in the opposite direction.  This is how the hard drive creates binary data.  Reading the data is done in a similar fashion.  Once the particles have been oriented in a certain way, the head will move over them, but without any electrical charge in the coil.  The head will then magnetize according to how the particle is oriented on the drive.  An electrical signal is then generated on the coil, and is sent down the arm and back to the drive where it is interpreted as either a 1 or a 0.

When the hard drive writes a file to the platters, it will calculate the amount of sectors required to save the file.  If the file is under 512bytes in size, then the hard drive will only require 1 sector.  If it is more, it will need more sectors.  If the file is 768 bytes in size - meaning 1.5 times the size of one sector - then the drive will use 2 sectors for the file.  The drive will save 512bytes to the first sector, filling it up, and then will save 256 bytes to the second sector, using exactly half of the sector's capacity (512/2=256).  The remaining 256bytes on the second sector cannot be used to save other files.  In this case, it becomes "wasted space".  The reason for this is because on the first tracks of the hard drive, there is a special part called the 'file allocation table' which is essentially a directory of all files contained on the drive.  When you write a file to the disk, the drive's head will first go to the file allocation table and find the first available sector on the drive.  It will then go to that sector and start writing the information.  When the sector is filled up, the head returns to the file allocation table (FAT) to find the next available sector.  Once the file is written down, the head will then mark in the FAT which sectors contain the file.  So the next time you go looking for it, the head will move over FAT, and find the sectors to which the file is saved onto, and then move to those sectors to retrieve the file.  Sometimes, files can be saved on sectors on different tracks and even on different platters.  When this happens, we call that data 'fragmented' because the file is spread all over the drive.  When a file is spread out like that, it causes the drive's head to have to move to different areas of the platters to find the data it needs, thus increasing seek times, meaning you will have to wait longer while the hard drive retrieves the file.

It is important to note that while the basic principle of FAT is the same for any drive, on any computer, using any Operating System, the technology of FAT is reserved to the Windows Operating System.  The first version of FAT was originally released with MS DOS and Windows 3.x.  The first version of FAT was called FAT16.  FAT16 had a serious limitation associated with it.  The biggest being partition sizes were limited to a maximum of 4gb, meaning if you had a hard drive larger than 4gb, it had to be partitioned into two or more drives.  Partitioning a drive means to take one physical hard drive and turning it into two or more logical drives.  Partitioning can be done with a utility called FDISK or when installing a distro of Linux.  When installing Windows, the option to partition the drive is not available.  The successor to FAT16 is what is most commonly used today, FAT32.  FAT32 was introduced with Windows 95, and carries through until Windows XP.  IT contained many improvements such as larger drive capacities.  The disadvantage is that it is not backwards compatible with MS DOS and Windows 3.x Operating Systems.  And you cannot view the files stored in a FAT16 partition.  With the introduction of Windows XP, a new file system had to be conceived.  In comes the New Technology File System, or NTFS.  Earlier versions of Windows cannot read files on NTFS partitions, but NTFS partitions can read earlier versions of FAT32.

After a hard drive has been partitioned, it must be formatted before it can be used.  The process of formatting a hard drive involves scanning the surface for any physical damage that cannot contain information.  During the format process any damaged areas of the disk and their surrounding sectors are marked as bad sectors and cannot be used.  At the end of the format, a report is displayed stating the number of bad sectors found.  Formatting can either be low-level or high-level.  This distinction is similar to when you are erasing a CD-RW with either quick erase or full erase.  A high-level format is the quickest way to format a drive, and is the most common.  When you are booting your computer from DOS and you type in 'format c:' at the prompt, this is a high-level format.  What it does is erase the data in the file allocation table (FAT), and writes it as having all sectors available for new data.  This allows the computer to display the hard drive as being empty.  However, this is not true.  All your data is still physically there, however the computer can no longer access it due to the loss of the FAT.  This can be dangerous if the hard drives contain sensitive or confidential information.  A low-level format erases everything on the disk.  It destroys the file allocation table and it re-writes all the tracks and sectors on the drive, giving them all 0s.  So instead of having a hard drive full of 1s and 0s which contain your data, you now have a hard drive with nothing but 0s on it.  This ensures any information on it is completely destroyed.  Low level formats are uncommon and take much longer to accomplish, but are much safer.

The platters and read/write heads inside the hard drive are sealed inside a metal case during manufacturing.  Hard drives are born inside clean rooms where the level of microscopic impurities is almost non-existent.  If any form of impurity, such as dust or hair got inside the drive, it would cause severe damage to the drive, including loss of data.  Each bit of data written to the disk is so small, that a spec of dust could easily shadow several tracks and sectors.  As well, dust could get stuck between the platter and the head, causing some scratching on the disk's surface, or worst, causing the head to crash.  When the head crashes, it collides with the platter, scratching the delicate surface and damaging the data.  This causes bad sectors, which are very often irreparable.  During a high-level format, the drive tries to fix bad sectors, but more often than not, it fails to do so.  Bad sectors are often the first sign of a dying hard drive.  The worst case scenario of a head crash is that the head for that side of the platter becomes damaged, which results in the loss of all your data on the platter, or worst, if that platter contains your FAT, then you loose all your data.  Special repair centers can still save the data, by physically opening the drive inside clean rooms and working on it, but their services are usually around 400$ per drive depending on the extent of the damage and the work that needs to be done.  So its a good idea to have backups of important files on CDs or DVDs or on a second drive.  Especially so if you are a small or large business.  This is less important for everyday users which can still live without their data.  Today's hard drives are extremely reliable and Seagate is now giving their drives 5 year warranties, but not for data.

After all that wonder that lies inside the actual hard drive, there is also some things to know about the outside of the hard drive.  Such as the printed circuit board (PCB) underneath the drive.  For a hard drive to work, it needs a controller.  When the computer sends data to be saved, it doesn’t tell the hard drive to save data to this sector on that track, and then move on to the sector on another track.  All the computer says is "Save this".  The 'save this' command is passed along to the hard drive controller which then takes the file to be saved and sees how large it is.  Then it instructs the drive to go to the file allocation table and find the first available sector to store the file.  The file is then sent, and saved onto the hard drive, and the location of the file is marked in the FAT.  In the early days of computing, before the AT spec of motherboard, ISA (Industry Standard Architecture) and PCI (Peripheral Component Interconnect) controller cards had to be installed so that the hard drive plugs into it instead of the motherboard.  This was way back when hard drives were a luxury item.  When computers didn’t come standard with one.  As the need for hard drives became more apparent, they were bundled with all computers.  And with this being done, it became impractical to have to always use a expansion card to control the hard drive.  This is why the controller soon became integrated into the hard drive itself, around the same time as the release of the AT spec.  The standard that came to be was called IDE, or Integrated Device Electronics.  The interface between the drive and the controller was made via wires that went from the drive's head and spindle to the PCB underneath the drive.  From there, it connected to the motherboard thanks to a 40-pin, 40-conductor parallel ribbon cable.  As hard drives evolved, the 40-pin cable just couldn’t give the data transfer needed, and thus was born E-IDE drives, or Enhanced Integrated Device Electronics.  These drives still ran on 40-pin cables, but instead of having 40 conductors, the E-IDE cables use 80 conductors.

IDE devices are connected to the motherboard via one of two channels, either primary or secondary.  Each channel can hold 2 drives maximum.  This allows most motherboards to hold a maximum of 4 hard drives unless and expansion card is used or unless the hard drive is equipped with RAID IDE channels.  More on RAID later.  To be able to have 2 hard drives per channel, there must be a way of identifying them.  This is done via what is known at the jumper.  A small series of pins next to the data cable on the hard drive with a 2-pin jumper which configures the drive as either master or slave.  If there are 2 devices on the same channel, one must be master and the other must be slave.  Which is which does not matter and does not affect performance in any way, it simply identifies the drives and creates a distinction between both of them.  Most hard drives have 4 settings possible on them.  Either Single Device, Master, Slave, and Cable Select.  Single Device is used mainly with Western Digital drives and is used if the hard drive is the only device on the channel.  Master and Slave denote the different drives on the same channel.  Cable Select is uncommonly used as it is often problematic.  Cable Select is used if another device is already on the channel, and you wish to add the drive to that channel, but you don’t want to configure it as master or slave.  During the POST (Power-On Self Test) when your computer boots up, it detects the drives on the channels.  If it detects a Single Device configuration, it stops there and proceeds to the second channel.  If it detects a master drive, it will then proceed to try and detect a slave drive.  If there is one, it will quickly detect it.  If there is none, it will try and try until it times out.  While this does not cause problems, it does add to the load time during boot up.  Same if there is only a slave drive without a master.  Cable select tells POST to first detect the other drive on the channel and see whether it is master or slave, and then it will configure the cable select drive as either master or slave accordingly.  Cable select cannot be used if there is only one drive on the channel, and you cannot have 2 cable select drives on the same channel.  As this would confuse POST.

While IDE and E-IDE are the evolution of the hard drive, the hard drive's controller has also evolved through the years.  The specification is called ATA, or Advanced Technology Attachment.  Now, the technology, when used with the older parallel cable, is called PATA, for Parallel ATA.  The ATA technology evolved through 6 phases, ATA1 to ATA6.

The new form of hard drive interface is called SATA, for Serial ATA.  Unlike parallel, which sends multiple bits of data at a time, Serial only sends one bit at a time.  While this may sound like a downgrade rather than an upgrade, the advantage of sending only one bit at a time is that the speed at which data is transferred can be far greater.  With parallel, you can only send the bits so fast before there starts to be corruption in the data stream.  SATA pushes the transfer speeds even higher than PATA did.  But that is not the only advantage.  SATA cables are also much smaller, having only 7 connections compared to the 40-pin of older PATA drives.  This makes SATA and PATA physically incompatible with each other.  An adapter can be used to connect your PATA drive with a SATA connection, but the opposite is not possible.  The final advantage SATA has over PATA is that the size of the cable makes for better internal airflow inside the case.  And if you have a window, SATA cables look so much prettier than the big bulky ribbon cables.  Due to this, parallel cables started coming out as round cables instead of ribbons.  What was done was to take the ribbon and cut each individual wire on the cable and group them together and sleeve the whole cable in either shrink-wrap or rubber tubing, thus creating a round cable which is slightly bigger than a SATA cable.  The disadvantage of this is that with all the wires grouped so close together, they become more susceptible to electromagnetic interference (EMI) which can cause the signal to degrade, thus leading to longer transfer times, and delays.  Another possible future use for SATA will be what is called 'hot swapping' meaning the ability to plug in a hard drive and remove it just as you would any other component, such as a memory stick, digital camera, MP3 player, without having to turn off the computer.

The final important component in the hard drive is what is called the cache.  The cache, just like the cache on the Central Processing Unit, is a small piece of memory which stores commonly used data.  Early hard drives didn’t have any cache, whereas most hard drives of today have 2mb.  All newer models now come with 8mb, while the industry standard still remains at 2mb.  When you retrieve data, the drive's head must move over FAT and locate the sectors to which the file is contained on, and then retrieve the actual data.  This is a long process, although it only takes a second or two.  Once the file has been retrieved, it is stored in the cache temporarily.  This way, if the data is called again, the hard drive simply has to load it from cache, and not go and retrieve it from the disk.  This cuts the time considerably, from seconds, to nanoseconds.  The effect of cache is most noticeable when you load up a program which must load several files, for example Valve Hammer Editor (VHE) is the level editor for Half-Life and Half-Life 2.  When VHE loads up, it must load several files, including textured, entities, help files, and other associated files.  The first time it loads up, it takes up to 10 seconds before everything is ready to go.  If you close the program, however, and immediately re-open it, you'll find the load time to be between one and three seconds.  This is because the data is still in the hard drive's cache.

Now that we know more about how the hard drive works and the basics of how the computer itself works, we can now learn more about the way data is stored.  Data is used throughout the computer as binary, and stored on the hard drive as such.  This is what is also called digital data.  Compare this to LPs and audio CDs and even audio tapes which store analog data.  Every one and zero is called a bit.  It is the smallest unit of measurement that is used.  Beyond that, we have the more common term, a byte.  A single byte is equivalent to 8 bits of data.  The reason for this is, at the moment, unknown.  Just like the reason for the placement of the keys on your keyboard.  Who knows why they put them like that.  Our number system is 10-base, meaning 1 to 10.  Binary data is 8-base, meaning 1 to 8.  We commonly say that 1000bytes is 1kilobyte, but it is not.  1kb is equivalent to 1024bytes, this due to the fact that 1byte is equal to 8bits.  This can often cause confusion, and this is the reason why a 80gb hard drive only contains 74gb.  Same with a 120gb HDD (Hard Drive Disk) which only really contains 111gb - Once again, 1000kb is often used to represent 1megabyte, when in fact it is 1024kb.  The next step after megabytes is gigabytes, a term which has been around for under 10 years.  And recently introduced to the public is the term terabyte, which signifies 1000gb, when in fact it is 1024.  This would mean that a 1tb hard drive is announced as having 10 000 000 000 000bits when in reality it contains 8 796 093 022 208bits.  This means that the announced size is larger than the actual size since hard drives or on 8-base numbering while we refer to them on the 10-base system.  This error is true for any form of data, but the computer industry is adapting to this anomaly slowly.  A 120gb hard drive has a warning label on it stating that it's true capacity is under 120gb, which is normal.  In our above example of a terabyte drive, you can see how much information must be contained on the platters.  Terabyte drives are typically four 250gb hard drives which are grouped together to form the terabyte.  This means that all those bits are split on about 12 to 16 platters.

In an effort to solve some of the common problems of hard drives, such as slow speeds, reliability, and adding more hard drive storage, a standard called RAID was created.  Redundant Array of Inexpensive Drives was first used with businesses, governments, and large corporations to assist them in one of many ways.  When RAID first came out, the most popular usage for it was for back up.  Several hard drives of the same capacity were bunched together and each stored the same data, so that if one hard drive failed, at least one other still contained the data.  This form of "dynamic" backup is what put an end to tape back up drives which were used earlier to back up data.  However, tape drives were slow, and didn't have as much room on them.  And it meant a technician had to be present while the back up was being done.  With RAID, as soon as the hard drive failed, all that was needed was a reboot, and the second drive would be used, and all the data would be persevered, thus doing away with roll backs.  This first form is called RAID-0, but is more commonly known as stripping.  The second form of RAID, called RAID-1 or more commonly, mirroring, was to add to disk size.  Back in the days, not only were hard drives failure prone, but they were also small.  You couldn't fit all the information in it that you wanted.  And adding a second hard drive meant having a second logical drive on your computer, such as drive C, drive D, drive E, and so on.  For large servers which needed lost of space, having 8 or more hard drives in it with each a different letter was impractical.  RAID does away with this by taking two or more physical drives and making them into 1 larger logical drive.  So if you had two 10gb hard drives in this RAID configuration, they would appear as only one 20gb hard drive.  If you have five 20gb hard drives, they would be viewed as one 100gb hard drive.  The final form of RAID, called RAID-?2? was introduced to increase the performance of the hard drive.  It is very similar to the first and second configuration, but it does not allow for backup, nor does it increase logical hard drive capacity.  This form of RAID takes the physical hard drives and views them as one logical drive, just like RAID-1, but does not add their hard drive space together.  It will write the same data to all drives on the array.  But unlike RAID-0, it cannot be used as backup.  The reason for this is that each disk in RAID-0 has a File Allocation Table on it, while in this RAID, only the first drive has a FAT.  So if the first drive fails, you loose all the data.  If any other drive fails, your data is still saved, but the array is lost.  RAID-?2? was used in servers where many clients requested information at the same time.  It was used to decrease read and write times.  While the first drive's head was writing to the FAT the location of the file, the second hard drive's head was already starting to save the file.  When the first drive's head was done writing to the FAT, it would then save the file.  But by that time the second drive would have already written it, and it would now be doing something else.  The same is true when data needs to be written.  And with more drives on the array, the faster the seek times could be.

However, with advances in technology, RAID is starting to become legacy technology.  Hard drives are becoming more and more reliable, which means that less and less businesses use RAID for backup.  Hard drives are also coming out in sizes of up to 250gb on a single drive, or using multiple drives in a RAID fashion, 1tb.  At the same time, server hard drives are being designed with rotational speeds of 10 000rpm, such as the Western Digital Raptor drives, and Seagate is releasing 15 000rpm server drives as well.  RAID is, despite these advances, still used.  Some organizations, such as Microsoft, governments, banks, and hospitals, cannot afford any kind of disturbance, so they utilize RAID.  For data backup of their databases.  Very large companies, such as Microsoft, Nortel, Corel, and universities and colleges use RAID to increase performance.  Even a 15 000rpm drive has limitations in speed.  But more importantly, these speed demons have serious limitations in capacity.  The Raptor drives from Western Digital are only available in 36gb and 74gb capacities, requiring the use of RAID-1 to combine drives for a larger logical drive.

The physical setup for RAID is quite simple, and can be done on any computer.  All that is needed is at least 2 hard drives, preferably of the same capacity and bonus if they are from the same manufacturer.  From there, you set them up just like you would normally set up your drives.  A two hard drive RAID configuration is sometimes called a 2-disk array.   If three hard drives were used, it would be called a 3-disk array, and so on.  The real magic of RAID lies in software.  A utility is configured to determine how your RAID array will work.




Other Forms of Data Storage

Data storage is a synonym to memory.  However, memory is reserved for temporary data storage such as RAM.  The term 'data storage' is used to permanent storage, such as the hard drive, floppy drives, optical drives, tape drives, and older methods, such as the punch card.  Since we have already covered the hard drive in the previous section, we wont be talking about it, and optical drives have their section after this one.  So this leaves us with talking about the floppy drives, tape drives, and other legacy devices.

Going in chronological order, the first form of data storage was the punch card.  The punch card has been in existence for several hundreds of years and were used even before computers of today came to be.  A punch card is a simple piece of paper cardboard with holes on it, which, when read by the computer, can be translated to either 1s or 0s depending on whether there is a hole or not.  The modern version of the punch card is the scantron, which uses black dots on the paper instead of holes.  This is commonly used during high school exams and voting.  For the punch card to work, it is past through a reader which takes the card and passes it under a series of LEDs or Light Emitting Diodes which will illuminate and, if there is a hole in the paper, the light will reach the other side where it will be picked up by a cell which will then charge up due to the light, and this will indicate a 1.  If there is no hole, the light cannot pass through, and the cell does not charge, indicating a 0.  Modern scantrons work in a similar fashion.  The role of the punch card as data storage has now become obsolete as the level of storage they can hold is very very small.  If you think about it, each hole represents one bit.  This would mean the possibility of storing even a megabyte of data impossible.  The scantron exists now however as a means of answering preset questions, instead of storing data, as mentioned, for school exams or regional voting where a black dot could indicate answer or candidate A and no dot could indicate answer or candidate B.

The next form of data storage that came out was the tape drive.  The name 'tape drive' does not referrer to the same kind of tape drive that is *somewhat* in use today as cheap data backup.  The original tape drive was like a very large audio tape.  Think of a large spindle containing a long ribbon covered with magnetic particles, just like a VHS tape or a music tape.  These spindles were mounted and used as data storage around the time of the PDP series of computers, in the 1970s or so.  This form of data storage was not RAM-type.  You could not access any part of the data at random.  The computer had to run past all the other data first, "fast forwarding" to the data it needs.  Today, this type of technology is being phased out by RAM-type devices such as CDs for music, DVDs for video, and, already replacing it for data storage is the hard drive.  But back then, the tape drive was a wonder.  For the first time, the computer had a form of permanent data storage with a large capacity.  This meant that programs could be saved to the tape drive instead of having the be re-inputed every time the computer was shut down.  The results of that said program could also be saved, instead of being lost.  The disadvantage was that retrieving the data and saving to the tape drive took considerable amount of time.  But in 1970 when tap drives were new and there was nothing like it before, people didn't mind it at all.

Following the tape drive very closely was the original floppy drive, the 5.25inch disks.  They were given the nickname 'floppy' drive because the drive media was flexible and could be bent to a certain extent without causing damage.  This was because the disk was made of a very thin sheet of ???? surrounded in a pliable plastic shell.  The disk was of a square shape with a slit at one end which allowed the actual disk to be removed from it's protective casing.  As well, a whole permitted the drive's read/write head to access the disk.  Unlike the modern 3.5inch floppy disks, the legacy floppy had no protective flap to protect the hole.  Most of the first home computers were equipped with one of these 5.25inch drives but without a hard drive, leaving the only method of data storage to be the floppy.  But with capacities of only about 520kb the amount of data that can be stored on one disk is very limited.  The evolution of the 5.25inch floppy drive was the 3.5inch floppy drive.  Although it was named floppy, the media was not flexible like its older brother.  If you tried to bend a disk, it would snap.  The 3.5inch floppy used a drive that fit in a 3.5inch bay instead of a 5.25inch bay, meaning the drive is roughly the same size as today's 3.5inch desktop hard drives.  Floppy drives, while about half the size as the older 5.25inch drives, could store up to 5 times the amount of data, to a maximum of 2.88mb uncompressed.  The standard, however, is 1.44mb.  Higher capacity can be achieved through BIOS settings for the drive and require a format of the disk.  That makes the disk unreadable to other drives that are not operating at the same data capacity.  The 3.5inch floppy also introduced a form of data protection by using a small tab on the bottom corner of the drive that could be set one way to make the disk 'read-only' or another way, to allow data on the disk to be changed.  While this wasn't a method of protecting data against willful damage, it did prevent accidental data loss.  When you insert the disk inside the drive, an LED lights up at the bottom corner.  If the tab is open, then a cell on the other side receives the light and charges itself sending a signal to allow data to be written to the disk.  If the corner is blocked via a tab, then the light from the LED does not pass through, and the data on the disk is then marked as 'read-only'.  3.5inch floppy drives were, up until a few years ago, very commonly used, even after more than 15 years of service to the computer industry.  However, now it is slowly being abandoned.  Many custom built computers provide the floppy drive as an "optional component" to the rest of the computer.  The reason for this is due to the higher capacity of CDs and their low costs.

Two other short lived forms of data storage were the ZIP drive and the Jazz drive.  The zip drive resembles the 3.5inch floppy drive in shape, but contains far more data, at around 100mb per disk.  ZIP drives rotate at a much higher speed than floppies, allowing faster access and read/write times than floppies.  Unlike floppy drives, the head for the ZIP drive does not actually touch the media.  This gives ZIP drives the ability to rotate faster.  Also, ZIP drives use the same technology that modern hard drives use to position its sectors, so that sectors on the outer tracks are the same size as those on the inner tracks.  The Jazz drive, was in fact a miniature hard drive inside a plastic casing.  A single platter was held in place inside a shell with a door on one end protected by a filter to prevent dust and impurities from touching the media's surface.  The drive's head would hover above the platter about a millimeter away, giving more head room than standard hard drives.  Jazz drives were very fragile and could only contain about 2gb of information on them, so they were really impractical.




Optical Drives

Yea




Video Cards

Being dumb as it is, the computer only understands 1s and 0s.  And to transform this to a display onto your monitor requires considerable amount of work.  But video cards were not always necessary.  The ENIAC did not have a video card.  The reason for this is that the video card transforms binary data into a visual display that we humans can interpret.  But when the ENIAC was busy churning away at calculations, it didn't display the results on a nice 17inch monitor.  It used a series of light bulbs to display binary data which was then translated by computer scientists.  But today, this would be impractical.  You cant play Half-Life 2 in binary data.  Nor can you type your English essay.  The first types of displays invented were simple monochrome monitors that could only display lines of text.  While displaying text may seem like a simple operation, it still requires quite a bit of processing power.  The first computers to use these types of monitors did not need a video card.  As the computer industry evolved, and with the creation of Windows, the need for a monitor that could handle more than just text was required.  Rendering the 2-dimensional Graphical User Interface (GUI) was too tasking on the processor, so the need for a secondary processor became reality.  This led to the invention of the 2D video card.  This card had it's own processor, then called the VPU for Visual Processing Unit.  It was somewhat like the processor but it only handled data relevant to displaying information on the monitor.  As monitors and Windows got more evolved, the need for better video cards increased again.  Cards now had to be able to display colour.  Eventually, computers started rendering 3-dimensional shapes.  This task was very heavy on the 2D cards, so the industry created 3D video cards.  Back then, a typical computer had two video cards, a 2D one and a 3D one.  Today, the 2D graphics card is obsolete seeing has how most computing requires the 3D card, but rendering 2D is still very common, so the 2D processing abilities of early 2D cards has been incorporated into 3D graphics cards.  These cards are now simply called video cards.  While their main function is the rendering of 3-dimensional environments, they also contain many features for rendering 2-dimensional work.  In the begging, many different companies existed that created 2D and 3D cards, but today, only two companies hold the thrown.  nVidia and ATI are the leading manufacturers of home and small business video cards.  Bleeding edge technology in the video domain remains with commercial products, such as with nVidia's series of Quadro cards and ATI's FireGL.  Another company involved in commercial products, and leading that market by far is 3Dlabs with their current product, the Wildcat Realizm 800.  This card is over 3 times better than current X800 and 6800 level video cards.  For example, the Yoda light saber scene in Star Wars: Episode 2 was rendered using the Wildcat Realizm 800 which has 640mb of onboard memory, compared to ATI and nVidia's finest which only have 256mb, and supports resolutions up to 3840*2400.

Today, video cards look almost like small motherboards.  A typical video card has it's own processor, it's own RAM, it's own I/O ports, and its own architecture, much like the motherboard has.  The difference is that video cards are much faster and much more complex.  Their Graphical Processing Units are even more complex than today's Pentium 4 processors, and contain more transistors.




Sound Cards

Sound is something we experience every minute of every day or our lives.  Yet, for most of us, it is something we do not understand.  How is sound produced?  How is it that I can hear?  While the goal is not to explain everything there is to know about sound, you need to understand how sound is produced and heard to understand fully how sound cards work.  The first sound devices on computers could only produce monotone sound, which was emitted from 8-bit speakers.  The same that produce the BIOS beeps in modern computers today are packed away at the bottom of your computer case, hidden from sight.  Now we have huge 7.1 speaker systems that literally bring the D-day scene from 'Saving Private Ryan' to life as you run up the beach, bullets buzzing past your ears, and the cries of fallen comrades.  Playing Counter-strike on a 4.1 makeshift speaker system allows me to hear where the enemy is, sometimes, before I can even see him.  Sound, although complicated to produce, is relatively easy for the computer to recreate.  The hardest part lies in, as mentioned, producing it.  Which is where the sound card usually comes in.  A computer works on binary data, which is called digital data.  Sound is not digital, it is analog.  Therefore, for a computer to produce sound, there must be a conversion.  When learning about sound and computers, it is now that learning the true difference between digital and analog become important.  The best example of digital and analog come from looking at a clock.  A digital clock changes from 20:56 to 20:57 abruptly, while an analog clock will have it's minute needle move fluidly from the 56 minute tick to the 57 minute tick.

So to understand how a soundcard works, we need to understand how sound works.  How sound is produced, and how it is understood by natural means, such as hearing the waves crash against the rock wall.  But to understand how we humans imitate and capture sound, with devices called speakers and microphones, we need to understand a little bit about electricity.  Sounds complicated doesnt it.  And yet, the sound card isnt as complicated as other devices such as the processor and motherboard.

Like ripples in a pond, sounds are merely vibrations.  Sound, light, radio waves, and computer frequencies are all part of a spectrum which can be described as a large wavelength.  At the lowest end of the spectrum, where the wavelength is very large, is sound.  This then moves to light, and then to Ultra Violet light, and x-rays and gamma rays, with computer clock frequencies in there as well.  Trying to understand and learn the spectrum could take one person their whole life.  We as humans know so little about ourselves, and our world.  In a very faint definition, sound and light are the same thing.  But we have eyes to see a part of the spectrum, ears to hear a part of the spectrum, vocal cords to produce parts of the spectrum and a brain to process all that information.  What is the difference?  Our ears are sensitive to certain levels of vibrations, and transform those vibrations into sound.  Our eyes are too, sensitive to much higher levels of vibration, which is what we call light.  People are much too often ignorant of these facts.  Anyways, I am not here to talk about phylosophy or my beliefs, or the way I view the world.  We are here to understand more about computers, and to expand our knowledge.  And in this case, we will learn a little bit more about sound.

When you talk, vocal cords in your larynx - the top portion of your trachea - vibrate together producing vibrations which are sound.  When a rock falls and hits the ground, it produces vibrations which our ears interpret as sound.  This can be quite confusing to most people and many spend their lives studying sound, and how it is produced and received by the human ear.  But, briefly, our ears contain canals filled with liquid, and an ear drum, which is a flap of skin which covers the inner ear, and is connected to nerves which send signals to our brain.  When a sound vibration enters our ear, it moves the ear drum, just like the speaker cone is moved by a magnet.  The nerves that are attached interpret the vibrations and relay them to our brain, which we then understand as sound.  This is why I can still hear my brother beating on his drumset even though he's in his room, door closed, and I am in mine, door closed.  Because vibrations are not completely limited by physical objects.

To be able to have a better understanding of how a speaker works, we need a basic knowledge of electricity.  You live with electricity every day, yet you dont know how it works.  All you know is, if I put my finger in the plug, it will hurt.  For the purpose of understanding electricity in use for speakers, we need to understand what is electricity, which means, protons and electrons.  It means atoms.  The building blocks of life.  So minute.  We humans live life at a grand scale.  Yet even a millimeter is a tremendous ammount of space.  It could hole thousands of transistors, or hundreds of biological cells, all of which are made with millions, if not more, of atoms.  Electricity uses positive and negatively charged atoms. Electricity has two main attributes, Amplitude, which is it's intensity, and Voltage, which is it's strength. 

If you take a look at a common speaker, you will find a magnet and two wires for electricity.  The front of the speaker is a shaped inwards just like a siphon, and in the middle of it, a small cone.  The whole front assembly is covered in a material which is stretched out.  This material, in combination with the cone, are what produce sound.  Electricity moves into the magnet, which then rapidly moves the speaker cone millimeters forward or backwards.  This then vibrates the whole speaker which then creates a sound.  Different voltages of electricity move the cone different amounts, producing different levels of sound.  This crude method mimics the way natural sounds are produced.  The same principle is used behind microphones.  A basic microphone uses a ribbon that is attached to electrical sensors.  When it captures a sound, say the voice of a signer, the ribbon vibrates and this sends electrical signals down the wires and back to a speaker, where those signals will move a speaker cone, thus projecting the signer's voice.

So now that we understand the very basics of sound, we are now ready to understand how a computer sound card works.  A sound card typically has a APU, or Audio Processing Unit to crunch the numbers instead of offloading the task to the CPU.  Most integrated sound motherboards don't have an APU, with the most common exception being the nForce2 chipset with the MCP-T audio accelerator.  The APU then sends it's data to a Digital-To-Analog Converter, which is a small chip on the soundcard.  The DAC chip transforms the digital binary data into different levels of electricity that will be sent off to the speaker to move the cone and produce different levels of sound.  The APU sends the sound levels as frequency numbers.  The DAC then reads each frequency and then changes it and outputs an electrical signal.  All soundcards today also allow the capture of sound, via either a microphone or a line-in, such as a CD player.  For this process, the digital to analog process is reversed.  An Analog-To-Digital Converter will measure each level of electrical current, and then assign a frequency based on that level.  That information is then sent to the APU for processing.  So as you can see, most soundcards have three main components, an APU, DAC, and ADC.




Expansion Cards

When IBM and Apple released their first computers, they did not come with any form of upgradability.  You couldn't add more RAM, could add a sound card, nor a video card, couldn't change the CPU.  All you could do was use it.  Once it became too old, you would have to buy a more recent model.  As competition between IBM and apple grew, they started to make expansion slots to which cards could be put in to give the computer more features, and to extend it's life.  The first type of expansion slot conceived was called Industry Standard Architecture, or simply ISA.  ISA slots are typically black and very long.  They were very common throughout the AT spec of motherboards but towards the end of the formfactor's life, they were becoming obsolete.  Early ATX motherboards had ISA slots for backwards compatibily with the older cards, so they usually only had one or two ISA slots on them.

The next type of expansion slot, which is still in common use today, is the Periphical Component Interconnect slot, or PCI.  This slot was much faster than the older ISA slot and was also physically smaller.  Typically white, these slots were introduced around the time of the AT formfactor.  When they started, there were only 2 or 3 on the motherboard, the rest of the slots being ISA.  But as the years went on, PCI overgrew ISA and drove it to extinction.  PCI slots, like ISA slots, are generic.  Anything can be connected to a PCI slot.  Your modem, sound card, video card, IDE controller card, NIC card, etc.  But as the computer and graphics industry evolved, a more powerful means of data communication between the video card and CPU were needed.

Thus came out AGP.  The brown Accelerated Graphics Port slot can be found on all ATX motherboards conceived until last year.  These slots are smaller than PCI slots, and there is only one.  It is at the top, near the CPU socket.  From there, you can connect your high end video card and it will have much faster communication with the CPU than a card of the same model would in a PCI slot.  This is because the AGP slot is connected to the northbridge, while the PCI slots are connected to the southbridge.

Up until a year ago, AGP was the slot of choice for your bleeding edge technology graphics card.  And for many gamers and computer users, it still is.  But for the world of graphics editing and digital movie making, the AGP slot had one major drawback.  While the transfer speed between it and the CPU was much faster than the PCI slot, it was still too slow for the demands of the video editing industry.  This lead to the birth of the PCI-E slot.  PCI-Express resembles the older PCI slot, and is used to most of today's high end video cards.  PCI-E is expected to replace not just AGP for video cards, but also replace older PCI slots, seeing as how the data transfer of PCI-E is much faster.  But due to the extreme popularity of PCI, you can expect it to take quite a few years before PCI becomes yesterday's tech.




The Power Supply Unit

yea




Cooling Solutions

Lets face it.  Computers are hot.  They generate a lot of heat, which must be dealt with in order to keep your computer from burning up.  Heat has been an issue ever since the very first computers.  ENIAC generated a lot of heat, due to it's vacuum tubes.  A cooling solution had to be designed.  But in the case of ENIAC, the cooling was required not so much for itself, but for the computer technicians who had to spend hours in the blistering hot room.  So, to cool down the room, fans were built in that would circulate the air and recycle it, thus bringing down the room temperature.

When the first PC's started appearing by IBM and Apple, they were a collection of circuits and transistors inside a metal box.  The same principle still in use today.  But they generated very little heat compared to today's monsters.  One of the secondary reasons why vacuum tubes were dropped for transistors.  Because transistors generate very little heat.  The first form of cooling found on computers was the power supply's fan.  The power supply has a job to do, and playing with electricity as such tends to generate a lot of heat.  So a fan was used to blow away the warm air into the room where it would mix with the cooler air.  For years, this was the only form of cooling inside the computer.  That is, until the later versions of the 486 and the first Pentium processors came out.  These CPUs generated heat.  This is because of the larger number of transistors they contained and the higher clock speeds of the chips.  The 486 only required what is called passive cooling.  Which means simply just a heatsink.  A heatsink is a piece of metal which rests on the component and has several fins which stick out into the air.  Heat from the CPU would be transfered to the heatsink, and then rise to the fins, where it would dissipate in the air.  The Pentium processor however, was too hot for just a simple heatsink.  It uses what is called active cooling.  Which generally means a heatsink and a fan.  A heatsink is once again placed onto the CPU, and at the top of the fins, a fan is attached which blow cool air onto the fins, thus helping with the transfer of heat.  This is commonly referred to as HSF which stands for heatsink and fan.  It is the most common form of cooling still in use today.

Over the years, more than just the CPU and power supply needed cooling.  As video cards became more powerful, their GPUs required cooling themselves.  Just like the CPU.  First they came with heatsinks, and later, heatsinks and fans.  Today, all high end video cards come with heatsink and fans which are almost as large as the card itself.  Another common component which needed cooling is the northbridge.  Since it is running at much higher speeds and must handle the traffic of the CPU, RAM, AGP or PCI-E video, it can get pretty warm.  Northbridge chips can use either passive or active cooling, depending on the motherboard manufacturer.  The final component that, in the last few years, has required cooling, is the case itself.  Its nice to have fans which blow away the warm air from the CPU and video card, but that air just gets blown into the case and is not recycled.  Case fans are simply just standard fans affixed to the front, rear, and sometimes side and top of the case in order to provide circulating air inside the case.  This lowers the temperature of all components, and allows the CPU fan to constantly blow cool air onto the heatsink.

Fans all work using electricity.  Some fans, you will find have only two wires, while others have three.  A two-wire fan has a red and black wire usually, and it is used for positive and negative respectively.  On a three-wire fan, the third wire is usually yellow or green, and sends a voltage signal back to the motherboard to give the speed of the fan, in revolutions per minute.

The driving mechanism within a fan is one of three different types, either ball bearing, sleeve bearing, or hydro bearing.  The most common being ball bearing, just like hard drive spindles, which have ball bearings within.  Ball bearing fans offer the longest life and best quality in a fan motor, but they are also a fair bit louder than the other forms.  Most ball bearing fans come with only one bearing, but more quality fans such as the Vantec Stealth, Thermaltake high performance fans, and others use two ball bearings.  This extra quality is always mentioned on the packaging as the manufacturer is proud of mentioning this.  Sleeve bearing fans are the cheaper versions of ball bearing fans.  They have a much shorter life expectancy but are more silent.  Another drawback to sleeve bearing fans is that they can only be mounted horizontally.  If you mount it vertically, such as a blow hole fan on the top of your case, the fan's mechanism will wear out faster and break, leading to premature failure of the fan.  Hydro.....

Fans come in many different shapes and sizes.  Standard fans - meaning the non fancy ones like the Thermaltake X-blower - come in sizes ranging from tiny to huge.  The size of a fan is always expressed in millimeters.  The smallest fans available, but not very common are the 20x20x5 fans.  These tiny things are 2 centimeters tall by 2 centimeters wide and half a centimeter deep.  They are not very popular at all.  The smallest common ones are the 40x40x10 fans which can be found most often in laptops and desktop hard drive coolers.  From then on we move to the 60x60x20 fans and the 80x80x10, 80x80x20 and the most common type, the 80x80x25 fans which are most used as case fans.  The x10 types mainly as small formfactor for CPU heatsinks and the x20 types are the second fan within the power supply unit.  After that we go on to the 92x92x25 and the 120x120x30 fans which, once again, are mainly used as case fans.

The size of the fan is inversely proportional to the noise they generate, in most circumstances.  A larger 120mm fan generates much less noise than a 40mm fan since the larger one can rotate slower to push the same amount of air.  However, not all fans are designed to push the same amount of air.  This is why the Vantec Stealth 80x80x25 case fans in my computer are much quieter than the 120x120x30 beast in my friend's case.  While he only has one exhaust fan, I have two.  This is so that I can achieve almost the same airflow as a single 120mm fan but by having two fans instead of only one, it permits me to buy quieter fans.  This is the same when considering fans of the same size.  The noise and airflow are opposites.  The Tornado from Vantec pushes about 4 times the amount of air than does the Stealth fan, but it is also louder by a factor of approximately 2.5 times.

Years ago, some people decided that using fans and air for cooling just wouldn't cut the mustard.  They needed a more powerful form of cooling.  This lead to the invention of water cooling.  Like radiator fluid in your car, water cooling runs water over the heat critical components inside the computer.  This is done with what is called water blocks, which are small plastic blocks to which you position over the CPU, video card, hard drives, and any other components.  Each block is of a specific shape and has a specific purpose.  You cannot place the CPU block on the hard drive and vice versa.  Tubing connects all the blocks together, and all of that to a reservoir located usually at the bottom of the case.  The reservoir acts like the radiator in a car.  A fan cools down the water so that it may then be pumped back into the system.  Most water cooling kits combine the reservoir and pump into one unit.  Therefore, a water cooled computer only has one fan and is much quieter than using air cooling.  Water cooling is, however, much more expensive, a lot more dangerous, due to the proximity of water – a simple tear in any pipe or block could send water spraying into your case, leading to dollars of damage – and it is why when you buy a water kit, it is highly recommended that you build it outside your case and test it before putting it inside so that if there happens to be a leak, it wont destroy your computer.  Water cooling is also less compatible and harder to install than fans, and needs regular maintenance.  A water cooling system needs to have it's water changed every 6 months or so to prevent bacterial growth, and needs special chemicals put in to help prevent the growth of bacteria.  The nice part at least, is that the chemical used is usually green and makes for a nice looking effect.  Water cooling, like air cooling, cannot reduce the temperature bellow room temperature.

Very powerful computers, or very rich enthusiasts use phase change cooling.  It is the same principle as your refrigerator, so you can see that it is not entirely new technology.  Phase change uses no fans, no liquids, and no heatsinks or water blocks.  It is therefore the smallest system, and is completely quiet.  A special adapter is placed over the CPU which creates a seal around it, and then air is sent over it, and ......




Assembling a Computer

Before you go out and buy all your different parts, it is important that you decide what your new computer will be used for.  Will it be a media and gamming server or simply used for homework, or anything in between.  And base your purchase on that.  From there, you need to decide whether you want to go with an Intel based system or an AMD based system.  Then you pick out a motherboard.  From there, you have yourself limited and can buy the rest from there.

This section is to help you assemble the computer, and ends at that.  I will not tell you which parts to buy for what, or how to install your Operating System or anything else.  This is a hardware paper, and my expert knowledge resides in computer hardware.  As well, I myself nor the websites directly and indirectly involved with this document can be held responsible for any damage caused to your equipment due to following this guide, or due to your own negligence.

First thing's first.  Before you take out all your new components, you have to prepare your working environment.  To assemble your computer you will need a phillips screw driver, this guide, and nothing else.  I have read many guides to PC building and they all start by telling you about static electricity.  Well, static electricity, comonly known as ESD for Electro-Static Discharges do in fact kill computers.  So it is important to minimize the risk.  This is why all your components come packaged in a black-ish plastic wrap called an anti-static bag.  Dont rip these, and keep them in a safe place.  Something all guides recommend it to build your computer on hard floors, and not on a carpet.  While still working on a table or bench of some sorts, unless you fancy working on the floor.  Hey, I do.  But I can tell you this, building a computer on a carpet floor is not suicide.  This is how I work, in my bedroom on my carpet floor.  I know full well what I am doing, the risks I am taking, and I am a daring person.  So for legalities, dont build on a carpet floor.  But know this, your not walking into certain death by doing so.

The first thing you'll want to do is take your case out of it's box.  Once it's out, unscrew both side panels, and keep the screws seperate with the side panels.  Most side panel screws do not have teeth under the head of the screw so as to not scratch the case and the paint.  Once the side panels are removed, you'll find inside a few cables which hang loosly and are connected to the front of the case.  These are often devided into three groups.  The first is the front I/O such as power, reset, HDD light, power light.  The second is usually by itself or grouped with the front I/O and is the speaker.  The final group would be for any features you have in the front, such as USB and sound jacks.  These can either be in block pins or in individual pins.  Inside the case you should also find a box or a bag containing screws, and micelanious other parts.  Remove the box.  Some cases feature motherboard trays that can be removed from the case itself to help in the installation of the motherboard.  If this is the situation with your case, remove the tray and set it flat on your work area.  Now pull your motherboard out of the box and out of it's anti-static bag.  While you can hold the motherboard anywhere, I myself choose to hold it by the edges to keep the motherboard surface pure, and clean of oil from our fingers.  Your motherboard will still work if you touch it.  Now hold your mobo next to your motherboard tray or case and match the screw holes on the board with the ???? on the case.  They should match exactly without having extra ???? on the case or without having holes without ????s.  Once that is done, you can now screw the motherboard into place.  Most cases comes with little red cardboard washers that you can use to protect the motherboard, but this is not really neccesary as today's motherboards have metal surrouding the screw holes for protection.

Once the motherboard is secured, you'll want to grab your processor from it's packaging and insert it into the motheboard.  Be very careful however, as processors can easily be damaged.  The bottom side contains hundreds of tiny pins which connect it to the motheboard.  If you bend a pin, you can often bend it back safely.  But if you shear off a pin, then you can kiss that 200$ processor good bye.  And dont think about refund, RMA (Return Manufacturer Authorisation), or warranty as they dont cover installation by yourself.  The CPU's pins are also arranged so that the processor can only fit the motherboard a certain way.  And different processors have different pin arrangements and different numbers of pins.  The top part of the CPU has little arrows and dots to indicate which way the processor goes in.  To insert it, you have to lift the locking arm on the side of the socket.  This arm is either plastic or metal.  It isnt a quality thing, as I myself preffer the plastic arm as it is easier to handle.  Lift the arm until it is perpendicular to the board, at a 90 degree angle.  As you lift it, you will see the top part of the socket move.  Once it is fully extended, the arm should lock.  You can then drop the processor into the socket, but dont take me literraly on that.  Dont drop it.  It should fit in without any force.  If you have to force, stop right there.  Once the processor is in, gently press down on it to confirm it is in proprelly, and then close the locking arm back to it's original position.

Intel's new LGA 775 Pentium 4 processor is a little different.  While I have no pictures to show you, I have seen them in real life.  The processor does not have pins, but instead holes.  The motherboard on the other hand, as the pins.  The concept is the same.  Raise the locking arm until it is perpendicular.  Then, you will have to raise a lid that covers the socket.  A piece of plastic protects the pins during shipment.  Remove this, and insert the processor.  Then close down the lid, and ensure the processor is in proprelly.  Then close down the locking arm.




Conclusion

I hope this document has been helpful in your quest to understanding just how computers work.  The computer is a very wonderful thing.  But understanding how a computer works, requires one to open his or her mind to things that we cannot comprehend on a human scale, such as integrated circuit chips.  And through my teachings of this document, it has allowed me to refresh my memory on certain things as well as to confirm my knowledge.  I have also learned a few things myself.  And through the writing of this document during long hours on my laptop, it allowed me to concentrate, and I realise now how ignorant the human species has become.  There is so little that we do know, and that we do understand.  Some people think they know everything about a computer, yet they do not know what is sound, they dont know how Random Access Memory works, or how the GPU turns millions of 1s and 0s into a beautiful image on your 21inch display.  I myself do not know everything there is to know about computers, and it would take me my life to know it all, if that.  But I can safely say that I know a lot about them, and a lot more than the typical individual who cant see past the Windows Graphical User Interface.

Writing this document has been a challenge, to which I stood proudly, knowing the work I had a head of me.  Above all though, writing this has given me a new outlook on life.  I am trying to be less ignorant, and accept that there is very little that I do know, and that most of what I know, like the iceberg, is only the tip, humans are only scratching the surface of science and understanding.  Many discoveries are yet to be made, and many scientists are yet to receive fame for their discoveries.  And I believe that now that you have a better understanding of how computers work, you will also have a much higher respect for them, and patience for them, as well as the manufacturers of each individual part.  It took humans hundreds of years to reach the level of understanding and knowledge that we today pocess.  Imagine what we will know tomorrow.  A hundred years ago, children were not learning mathmatics in elementary school, and trigonometry and physics were materials covered in the most recognised universities, instead of common grade 10 classes each child passes through.  Image tomorrow.  Spacial reasonning, physics, biology, quantom physics, and much more we have even yet to dream of.  We know a lot.  But the more you know, the more you realise you dont know.  What is time?  What are dreams?  What is communication..  Humans are described as having 5 sences, sight, smell, touch, taste, and hearing.  But in the near future others may join, such as sence.  Have you ever noticed that if you stare at someone on the bus, they will turn and look at you after a few seconds?  Inexplicable isnt it.  Inexplicable today.  Tomorrow, we might call it the sixth sence.  There is so little we know and even more little that we actually understand.

I have challenged myself to understand more about computers in the writing of this document, and now that I look back at the many hours spent writing it, I realise I have learned a lot more than just 'how computers work'.  I have become enlightened.  To which I encourage all of you to challenged yourselves and learn.  Discover, learn,  understand, teach.  These are the four steps to which human evolutions abide by.  We discover something, learn about it, understand how it works, how it has come to be, and then we teach our fellow humans, and our children, so that we may expand our knowledge, and evolve.  Expand your knowledge, and grow.  Grow as a person, as an individual.

You may ask how I came to have all this knowledge.  No I was not born a genious, and no I was not born with parents who lived with computers, or were computer-literate.  My parents bought their first computer in 1992 as a family computer for the main reasons of work, and because at the time it was the “hype”.  Everyone needed one.  Our 486 had many problems and we often sent it to a personal business for repair, maintenance, and upgrade.  I was, however, born curious, and perseverant.  I wondered how it is that a big beige box worked to play Cosmos and Commander Keep and how it is some Chinesse guy, your average father of 3 girls, was able to fix any problem we had.  I wondered why it is everytime we sent it to him, he often had to erase the data that was in it, such as my save games to Duke Nukem 3D and to Tie Fighters.  A few years later, sure enough, I was taking apart the old computer to see what was inside.  I didnt know anything.  I didnt know what a processor was or what was a hard drive.  But I played around and explored.  And as you can guess, I took it appart, and it was only 2 years later that I gained sufficient knowledge to be able to put it back together again.  Since then my knowledge has increased by many factors.  I now have my own desktop computer which I have built myself, and a laptop which I have striped to guts and rebuild.  I disassemble my desktop monthly to clean it, and I take care of it.  It is my life.

How did I grow up?  While other kids played with toy guns or beat up kids in the school yard, I was at home.  I was doing what every child has done at least once.  I was building something.  My childhood consisted of a black tool box filled to the brim with Legos – sadly, now owned my Mega Blocks – and I would build all sorts of buildings, and cities, and vehicles.  I would then go to a friend's house and we would play together, for hours on end.  I wouldnt buy a kit and follow the instructions though.  At first yes, but then I would all mix my pieces together and create my own models, my own sets.  And I would dream of stories and advantures that my legos would lie through.  I was not spoiled by my parents, and I was not deprived or poor either.

I just seeked better understanding, I am curious, and perseverant.  This is what has driven me from birth, and it is what will drive me until death.  Again, I invite you all to chalenges yourselves.  Discover – Learn – Understand – Teach.

Pilot