Hard Drive Technologies

<< Semiconductor Memory vs Hard Disk, Mechanical Delays and Flash Memory

Arithmetic Logic Shift Unit - ALSU, Radix Conversion, Fixed Point Numbers >>

"It would appear that we have

reached the limits of what it is

possible to achieve with

computer technology, although

one should be careful with such

statements, as they tend to

sound pretty silly in five years."

--JOHN VON NEUMANN, 1949

In this chapter, you will learn

f all the hardware on a PC, none gets more attention--or gives more

how to

anguish--than the hard drive. There's a good reason for this: if the hard

Explain how hard drives work

■

drive breaks, you lose data. As you probably know, when the data goes, you

Identify and explain the ATA

■

have to redo work or restore from backup--or worse. It's good to worry about

hard drive interfaces

the data, because the data runs the office, maintains the payrolls, and stores the

Identify and explain the SCSI

■

hard drive interfaces

e-mail. This level of concern is so strong that even the most neophyte PC users

Describe how to protect data

■

are exposed to terms such as IDE, ATA, and controller--even if they don't put

with RAID

the terms into practice!

Explain how to install drives

■

This chapter focuses on how hard drives work, beginning with the internal

Configure CMOS and install

■

drivers

layout and organization of the hard drive. You'll look at the different types of

Troubleshoot hard drive

■

hard drives used today (PATA, SATA, and SCSI), how they interface with the

installation

PC, and how to install them properly into a system. The chapter covers how

more than one drive may work with other drives to provide data safety and

improve speed through a feature called RAID. Let's get started.

101

How Hard Drives Work

■

All hard drives are composed of individual disks, or platters, with read/

write heads on actuator arms controlled by a servo motor--all contained in

a sealed case that prevents contamination by outside air (see Figure 8.1).

The aluminum platters are

coated with a magnetic medium.

Cross Check

Two tiny read/write heads ser-

vice each platter, one to read the

Implementing Hard Drives

top and the other to read the bot-

In the Essentials book you covered the process of implementing a

tom of the platter (see Figure 8.2).

preinstalled drive. What are the two steps that must be performed on ev-

The coating on the platters is

ery installed hard drive so that the operating system can use the drive?

phenomenally smooth! It has to

be, as the read/write heads actu-

ally float on a cushion of air above

the platters, which spin at speeds between 3500 and 10,000

rpm. The distance (flying height) between the heads and the

disk surface is less than the thickness of a fingerprint. The

closer the read/write heads are to the platter, the more

densely the data packs onto the drive. These infinitesimal

tolerances demand that the platters never be exposed to out-

side air. Even a tiny dust particle on a platter would act like

a mountain in the way of the read/write heads and would

cause catastrophic damage to the drive. To keep the air clean

inside the drive, all hard drives use a tiny, heavily filtered

aperture to keep the air pressure equalized between the

interior and the exterior of the drive.

Data Encoding

Although the hard drive stores data in binary form, visual-

izing a magnetized spot representing a one and a non-mag-

netized spot representing a zero grossly oversimplifies

the process. Hard drives store data in tiny magnetic

· Figure 8.1

Inside the hard drive

· Figure 8.2

Top and bottom read/write heads and armatures

102

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

fields--think of them as tiny magnets

that can be placed in either direction on

the platter, as shown in Figure 8.3. Each

tiny magnetic field, called a flux, can · Figure 8.3 Data is stored in tiny magnetic fields.

switch back and forth through a process

called a flux reversal (see Figure 8.4).

Electronic equipment can read and

write flux reversals much faster and

easier than it can magnetize or not mag-

netize a spot to store a one or a zero.

In early hard drives, as the read/

write head moved over a spot, the direc-

tion of the flux reversal defined a one or

a zero. As the read/write head passed

from the left to the right, it recognized · Figure 8.4 Flux reversals

fluxes in one direction as a zero and the

other direction as a one (Figure 8.5).

Hard drives read these flux reversals at

a very high speed when accessing or

writing data.

Today's hard drives use a more

complex and efficient method to

· Figure 8.5 Fluxes are read in one direction as 0 and the other direction as 1.

interpret flux reversals using special

data encoding systems. Instead of

reading individual fluxes, a modern

hard drive reads groups of fluxes called

runs. Starting around 1991, hard drives

began using a data encoding system

known as run length limited (RLL). With

RLL, any combination of ones and

zeroes can be stored in a preset

combination of about 15 different runs.

The hard drive looks for these runs and

reads them as a group, resulting in

much faster and much more dense data.

Whenever you see RLL, you also see

two numbers: the minimum and the

· Figure 8.6 Sequential RLL runs

maximum run length, such as RLL 1,7

or RLL 2,7. Figure 8.6 shows two

sequential RLL runs.

Current drives use an extremely advanced method of RLL called Par-

tial Response Maximum Likelihood (PRML) encoding. As hard drives pack

more and more fluxes on the drive, the individual fluxes start to interact

with each other, making it more and more difficult for the drive to verify

where one flux stops and another starts. PRML uses powerful, intelligent

circuitry to analyze each flux reversal and to make a "best guess" as to what

type of flux reversal it just read. As a result, the maximum run length for

PRML drives reaches up to around 16 to 20 fluxes, far more than the 7 or so

on RLL drives. Longer run lengths enable the hard drive to use more com-

plicated run combinations so that the hard drive can store a phenomenal

103

Chapter 8: Hard Drive Technologies

amount of data. For example, a run of only 12 fluxes on a hard drive might

equal a string of 30 or 40 ones and zeroes when handed to the system from

the hard drive.

The size required by each magnetic flux on a hard drive has reduced

considerably over the years, resulting in higher capacities. As fluxes become

smaller, they begin to interfere with each other in weird ways. I have to say

weird since to make sense of what's going on at this subatomic level (I told

you these fluxes were small!) would require that you take a semester of

quantum mechanics! Let's just say that laying fluxes flat against the platter

has reached its limit. To get around this problem, hard drive makers re-

cently began to make hard drives that store their flux reversals vertically (up

and down) rather than longitudinally (forward and backward), enabling

them to make hard drives in the 1 terabyte (1024 gigabyte) range. Manufac-

turers call this vertical storage method perpendicular recording (Figure 8.7).

For all this discussion and detail on data encoding, the day-to-day PC

technician never deals with encoding. Sometimes, however, knowing what

you don't need to know helps as much as knowing what you do need to

know. Fortunately, data encoding is inherent to the hard drive and com-

pletely invisible to the system. You're never going to have to deal with data

encoding, but you'll sure sound smart when talking to other PC techs if you

know your RLL from your PRML!

Moving the Arms

The read/write heads move across the platter on the ends of actuator arms. In

the entire history of hard drives, manufacturers have used only two technol-

ogies to move the arms: stepper motor and voice coil . Hard drives first used

stepper motor technology, but today they've all moved to voice coil.

Stepper motor technology moved the arm in fixed increments or steps,

but the technology had several limitations that doomed it. Because the inter-

face between motor and actuator arm required minimal slippage to ensure

precise and reproducible movements, the positioning of the arms became

less precise over time. This physical deterioration caused data transfer er-

Floppy disk drives still use

rors. Additionally, heat deformation wreaked havoc with stepper motor

stepper motors.

drives. Just as valve clearances in automobile engines change with operat-

ing temperature, the positioning accuracy changed as the PC operated and

various hard drive components got warmer. Although very small, these

changes caused problems. Accessing the data written on a cold hard drive,

for example, became difficult after the disk warmed. In addition, the read/

write heads could damage the disk surface if not "parked" (set in a non-data

area) when not in use, requiring techs to use special parking programs

before transporting a step-

per motor drive.

All hard drives made

today employ a linear motor

to move the actuator arms.

The linear motor, more

popularly called a voice coil

motor, uses a permanent

magnet surrounding a coil on

· Figure 8.7 Perpendicular versus traditional longitudinal recording

the actuator arm. When an

104

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

electrical current passes, the coil generates a magnetic field that moves the

Tech Tip

actuator arm. The direction of the actuator arm's movement depends on the

polarity of the electrical current through the coil. Because the voice coil and the

Fluid Bearings

actuator arm never touch, no degradation in positional accuracy takes place

Currently, almost all hard drives

over time. Voice coil drives automatically park the heads when the drive loses

use a motor located in the center

power, making the old stepper motor park programs obsolete.

spindle supporting the drive plat-

ters. Traditionally, tiny ball bear-

Lacking the discrete "steps" of the stepper motor drive, a voice coil drive

ings support the spindle motor,

cannot accurately predict the movement of the heads across the disk. To

and as disk technology has ad-

make sure voice coil drives land exactly in the correct area, the drive re-

vanced, these ball bearings have

serves one side of one platter for navigational purposes. This area essen-

become the limiting factor in the

tially "maps" the exact location of the data on the drive. The voice coil

three critical design criteria for

moves the read/write head to its best guess about the correct position on the

hard drives: rotational speed,

hard drive. The read/write head then uses this map to fine-tune its true po-

storage capacity, and noise levels.

sition and make any necessary adjustments.

The higher the rotational speed of

Now that you have a basic understanding of how a drive physically

a drive, the more the metal-on-

stores data, let's turn to how the hard drive organizes that data so we can

metal contact creates heat and lu-

use that drive.

bricant problems that impact the

lifespan of the bearings. However

precisely machined, ball bearings

Geometry

are not perfectly round. The mea-

surement of how much they wob-

Have you ever seen a cassette tape? If you look at the actual brown Mylar (a

ble (and thus how much the drive

type of plastic) tape, nothing will tell you whether sound is recorded on that

platters wobble), called runout, is

tape. Assuming the tape is not blank, however, you know something is on the

now the limiting factor on how

tape. Cassettes store music in distinct magnetized lines. You could say that

densely you can pack information

the physical placement of those lines of magnetism is the tape's "geometry."

together on a disk drive.

Geometry also determines where a hard drive stores data. As with a cas-

The technological fix for this

sette tape, if you opened up a hard drive, you would not see the geometry.

comes in the form of fluid bear-

But rest assured that the drive has geometry; in fact, every model of hard

ings. A fluid bearing is basically

a small amount of lubricant

drive uses a different geometry. We describe the geometry for a particular

trapped in a carefully machined

hard drive with a set of numbers representing three values: heads, cylin-

housing. The use of fluid in place

ders, and sectors per track.

of metal balls means that no con-

tact occurs between metal sur-

Heads

faces to generate heat and wear.

The number of heads for a specific hard drive describes, rather logically, the

The fluid also creates no mechani-

number of read/write heads used by the drive to store data. Every platter

cal vibration, so fluid bearings

can support higher rotational

requires two heads. If a hard drive has four platters, for example, it would

speeds. The runout of a fluid

need eight heads (see Figure 8.8).

bearing is about one-tenth that of

Based on this description of heads, you would think that hard drives

the best ball bearing, significantly

would always have an even number of heads, right? Wrong! Most hard

increasing potential information

drives reserve a head or two for their own use. Therefore, a hard drive can

density. The absence of a mechan-

have either an even or an odd number of heads.

ical connection between moving

parts also dramatically reduces

Cylinders

noise levels, and the fluid itself

acts to dampen the sound further.

To visualize cylinders, imagine taking a soup can and opening both ends of

Finally, liquid bearings provide

the can. Wash off the label and clean out the inside. Now look at the shape of

better shock resistance than ball

the can; it is a geometric shape called a cylinder. Now imagine taking that

bearings.

cylinder and sharpening one end so that it easily cuts through the hardest

metal. Visualize placing the ex-soup can over the hard drive and pushing it

down through the drive. The can cuts into one side and out the other of

105

Chapter 8: Hard Drive Technologies

· Figure 8.8

Two heads per platter

each platter. Each circle transcribed by the can is where you store data on the

drive, and is called a track (Figure 8.9).

Each side of each platter contains tens of thousands of tracks. Interest-

ingly enough, the individual tracks themselves are not directly part of the

drive geometry. Our interest lies only in the groups of tracks of the same di-

ameter, going all of the way through the drive. Each group of tracks of the

same diameter is called a cylinder (see Figure 8.10).

There's more than one cylinder! Go get yourself about a thousand more

cans, each one a different diameter, and push them through the hard drive.

A typical hard drive contains thousands of cylinders.

Sectors per Track

Now imagine cutting the hard drive like a birthday cake, slicing all the

tracks into tens of thousands of small slivers. Each sliver is called a sector,

and each sector stores 512 bytes of data (see Figure 8.11). Note that sector

· Figure 8.9

Tracks

106

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

refers to the sliver when discussing geometry, but it refers to the

specific spot on a single track within that sliver when discussing

data capacity.

The sector is the universal "atom" of all hard drives. You

can't divide data into anything smaller than a sector. Although

sectors are important, the number of sectors is not a geometry.

The geometry value is called sectors per track (sectors/track).

The sectors/track value describes the number of sectors in each

track (see Figure 8.12).

The Big Three

Cylinders, heads, and sectors/track combine to define the hard

drive's geometry. In most cases, these three critical values are

referred to as CHS. The importance of these three values lies in

the fact that the PC's BIOS needs to know the drive's geometry

to know how to talk to the drive. Back in the old days, a techni-

cian needed to enter these values into the CMOS setup program

manually. Today, every hard drive stores the CHS information

in the drive itself, in an electronic format that enables the BIOS

to query the drive automatically to determine these values.

You'll see more on this later in the chapter in the section called

"Autodetection."

Two other values--write precompensation cylinder and · Figure 8.10

Cylinder

landing zone--no longer have relevance in today's PCs; how-

ever, these terms still are tossed around and a few CMOS setup

utilities still support them--another classic example of a technology appen-

dix! Let's look at these two holdouts from another era so when you access

CMOS, you won't say, "What the heck are these?"

Write Precompensation Cylinder

Older hard drives had a real problem with the fact that sectors toward the

inside of the drives were much smaller than sectors toward the outside. To

handle this, an older drive would spread data a little farther apart once it got

to a particular cylinder. This cylinder was called the write precompensation

· Figure 8.11

Sectors

107

Chapter 8: Hard Drive Technologies

(write precomp) cylinder, and the PC had to know

which cylinder began this wider data spacing. Hard

drives no longer have this problem, making the write

precomp setting obsolete.

Landing Zone

On older hard drives with stepper motors, the landing

zone value designated an unused cylinder as a "parking

· Figure 8.12

place" for the read/write heads. As mentioned earlier,

Sectors per track

old stepper motor hard drives needed to have their

read/write heads parked before being moved in order to

avoid accidental damage. Today's voice coil drives park

themselves whenever they're not accessing data, automatically placing the

read/write heads on the landing zone. As a result, the BIOS no longer needs

the landing zone geometry.

IT Technician

Tech Tip

IDE

ATA--The King

■

The term IDE (integrated drive

electronics) refers to any hard

Over the years, many interfaces existed for hard drives, with such names as

drive with a built-in controller.

ST-506 and ESDI. Don't worry about what these abbreviations stood for;

All hard drives are technically

neither the CompTIA A+ certification exams nor the computer world at

IDE drives, although we only use

large has an interest in these prehistoric interfaces. Starting around 1990, an

the term IDE when discussing

interface called ATA appeared that now virtually monopolizes the hard

ATA drives.

drive market. ATA hard drives are often referred to as integrated drive elec-

tronics (IDE) drives. Only one other type of interface, the moderately popu-

lar small computer system interface (SCSI), has any relevance for hard

Tech Tip

drives. ATA drives come in two basic flavors. The older parallel ATA

(PATA) drives send data in parallel, on a 40- or 80-wire data cable. PATA

External Hard Drives

drives dominated the industry for more than a decade but are being re-

A quick trip to any major com-

placed by serial ATA (SATA) drives that send data in serial, using only one

puter store will reveal a thriving

wire for data transfers. The leap from PATA to SATA is only one of a large

trade in external hard drives. You

used to be able to find external

number of changes that have taken place over the years with ATA. To appre-

drives that connected to the slow

ciate these changes, we'll run through the many ATA standards forwarded

parallel port, but external drives

over the years.

today connect to a FireWire,

Hi-Speed USB 2.0, or eSATA

ATA-1

port. All three interfaces offer

high data transfer rates and

When IBM unveiled the 80286-powered IBM PC AT in the early 1980s, it in-

hot-swap capability, making them

troduced the first PC to include BIOS support for hard drives. This BIOS

ideal for transporting huge files

supported up to two physical drives, and each drive could be up to 504 MB--

such as digital video clips. Re-

far larger than the 5-MB and 10-MB drives of the time. Although having

gardless of the external interface,

however, inside the casing you'll

built-in support for hard drives certainly improved the power of the PC, at

find an ordinary PATA or SATA

that time, installing, configuring, and troubleshooting hard drives could at

drive, just like those described in

best be called difficult.

this chapter.

To address these problems, Western Digital and Compaq developed a

new hard drive interface and placed this specification before the American

108

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

National Standards Institute (ANSI) committees, which in turn put out the AT

Attachment (ATA) interface in March of 1989. The ATA interface specified a

The ANSI subcommittee di-

cable and a built-in controller on the drive itself. Most importantly, the ATA

rectly responsible for the ATA

standard is called Technical

standard used the existing AT BIOS on a PC, which meant that you didn't

Committee T13. If you want to

have to replace the old system BIOS to make the drive work--a very impor-

know what's happening with

tant consideration for compatibility but one that would later haunt ATA

ATA, check out the T13 Web site:

drives. The official name for the standard, ATA, never made it into the com-

www.t13.org.

mon vernacular until recently, and then only as PATA to distinguish it from

SATA drives.

Early ATA Physical Connections

The first ATA drives connected to the computer with a 40-pin ribbon cable

that plugged into the drive and to a hard drive controller. The cable has a

colored stripe down one side that denotes pin 1 and should connect to the

drive's pin 1 and to the controller's pin 1. Figure 8.13 shows the "business

end" of an early ATA drive, with the connectors for the ribbon cable and the

power cable.

The controller is the support circuitry that acts as the intermediary be-

tween the hard drive and the external data bus. Electronically, the setup

looks like Figure 8.14.

Wait a minute! If ATA drives

are IDE (see the Tech Tip), they al-

ready have a built-in controller.

Why do they then have to plug into

a controller on the motherboard?

Well, this is a great example of a

term that's not used properly, but

everyone (including the mother-

board and hard drive makers) uses

it this way. What we call the ATA

controller is really no more than an

interface providing a connection to

the rest of the PC system. When

your BIOS talks to the hard drive, it · Figure 8.13 Back of IDE drive showing 40-pin connector (left), jumpers (center), and

actually talks to the onboard cir-

power connector (right)

cuitry on the drive, not the connec-

tion on the motherboard. But, even

though the real controller resides on the hard drive, the

40-pin connection on the motherboard is called the con-

troller. We have a lot of misnomers to live with in the

ATA world!

The ATA-1 standard defined that no more than two

drives attach to a single IDE connector on a single ribbon

cable. Because up to two drives can attach to one connec-

tor via a single cable, you need to be able to identify each

drive on the cable. The ATA standard identifies the two

different drives as "master" and "slave." You set one

drive as master and one as slave using tiny jumpers on · Figure 8.14 Relation of drive, controller, and bus

the drives (Figure 8.15).

109

Chapter 8: Hard Drive Technologies

Directions

Jumper

· Figure 8.15

A typical hard drive with directions for setting a jumper

The controllers are on the motherboard and manifest themselves as two

40-pin male ports, as shown in Figure 8.16.

PIO and DMA Modes

If you're making a hard drive standard, you must define both the method

and the speed at which the data's going to move. ATA-1 defined two differ-

ent methods, the first using programmed I/O (PIO) addressing and the sec-

ond using direct memory access (DMA) mode.

· Figure 8.16

IDE interfaces on a motherboard

110

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

PIO is nothing more than the traditional I/O addressing scheme, where

the CPU talks directly to the hard drive via the BIOS to send and receive

data. Three different PIO speeds called PIO modes were initially adopted:

PIO mode 0: 3.3 MBps (megabytes per second)

■

PIO mode 1: 5.2 MBps

■

PIO mode 2: 8.3 MBps

■

DMA modes defined a method to enable the hard drives to talk to RAM

directly using old-style DMA commands. (The ATA folks called this single

word DMA.) This old-style DMA was slow, and the resulting three ATA sin-

gle word DMA modes were also slow:

Single word DMA mode 0: 2.1 MBps

■

Single word DMA mode 1: 4.2 MBps

■

Single word DMA mode 2: 8.3 MBps

■

When a computer booted up, the BIOS queried the hard drive to see

what modes it could use and would then automatically adjust to the fastest

mode.

ATA-2

In 1990, the industry adopted a series of improvements to the ATA standard

called ATA-2. Many people called these new features Enhanced IDE (EIDE) .

The terms ATA, IDE, and EIDE

EIDE was really no more than a marketing term invented by Western Digi-

are used interchangeably.

tal, but it caught on in common vernacular and is still used today, although

its use is fading. Regular IDE drives quickly disappeared, and by 1995, EIDE

drives dominated the PC world. Figure 8.17 shows a typical EIDE drive.

· Figure 8.17

EIDE drive

111

Chapter 8: Hard Drive Technologies

ATA-2 was the most important ATA standard, as it included powerful

new features such as higher capacities; support for nonhard drive storage

devices; support for two more ATA devices, for a maximum of four; and

substantially improved throughput.

Higher Capacity with LBA

IBM created the AT BIOS to support hard drives many years before IDE

Hard drive makers talk about

drives were invented, and every system had that BIOS. The developers of

hard drive capacities in millions

IDE made certain that the new drives would run from the same AT BIOS

and billions of bytes, not mega-

command set. With this capability, you could use the same CMOS and BIOS

bytes and gigabytes!

routines to talk to a much more advanced drive. Your motherboard or hard

drive controller wouldn't become instantly obsolete when you installed a

new hard drive.

Unfortunately, the BIOS routines for the original AT command set allowed

a hard drive size of only up to 528 million bytes (or 504 MB--remember that a

mega = 1,048,576, not 1,000,000). A drive could have no more than 1024 cyl-

inders, 16 heads, and 63 sectors/track:

1024 cylinders × 16 heads × 63 sectors/track × 512 bytes/sector = 504 MB

For years, this was not a problem. But when hard drives began to ap-

proach the 504 MB barrier, it became clear that there needed to be a way of

getting past 504 MB. The ATA-2 standard defined a way to get past this limit

with logical block addressing (LBA) . With LBA, the hard drive lies to the

computer about its geometry through an advanced type of sector transla-

tion. Let's take a moment to understand sector transla-

tion, and then come back to LBA.

Sector Translation Long before hard drives approached

the 504 MB limit, the limits of 1024 cylinders, 16 heads,

and 63 sectors/track gave hard drive makers fits. The

big problem was the heads. Remember that every two

heads means another platter, another physical disk that

you have to squeeze into a hard drive. If you wanted a

hard drive with the maximum number of 16 heads, you

would need a hard drive with eight physical platters

inside the drive! Nobody wanted that many platters: it

made the drives too tall, it took more power to spin up

the drive, and that many parts cost too much money (see

Figure 8.18).

· Figure 8.18

Too many heads

Manufacturers could readily produce a hard drive

that had fewer heads and more cylinders, but the stupid

1024/16/63 limit got in the way. Plus, the traditional

sector arrangement wasted a lot of useful space. Sectors

toward the inside of the drive, for example, are much

shorter than the sectors on the outside. The sectors on the

outside don't need to be that long, but with the tradi-

tional geometry setup, hard drive makers had no choice.

They could make a hard drive store a lot more informa-

tion, however, if hard drives could be made with more

· Figure 8.19

Multiple sectors/track

sectors/track on the outside tracks (see Figure 8.19).

112

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

The ATA specification was designed to have two geometries. The physi-

cal geometry defined the real layout of the CHS inside the drive. The logical

geometry described what the drive told the CMOS. In other words, the IDE

drive "lied" to the CMOS, thus side-stepping the artificial limits of the BIOS.

When data was being transferred to and from the drive, the onboard cir-

cuitry of the drive translated the logical geometry into the physical geome-

try. This function was, and still is, called sector translation .

Let's look at a couple of hypothetical examples in action. First, pretend

that Seagate came out with a new,

cheap, fast hard drive called the

ST108. To get the ST108 drive fast

Table 8.1

Seagate's ST108 Drive Geometry

and cheap, however, Seagate had

ST108 Physical

BIOS Limits

to use a rather strange geometry,

shown in Table 8.1.

Cylinders

2048

Cylinders

1024

Notice that the cylinder num-

Heads

ber is greater than 1024. To over-

come this problem, the IDE drive

Sectors/Track

performs a sector translation that

Total Capacity

108 MB

reports a geometry to the BIOS that

is totally different from the true ge-

ometry of the drive. Table 8.2

Table 8.2

Physical and Logical Geometry of the ST108 Drive

shows the actual geometry and the

"logical" geometry of our mythical

Physical

Logical

ST108 drive. Notice that the logical

Cylinders

2048

Cylinders

512

geometry is now within the accept-

able parameters of the BIOS limita-

Heads

tions. Sector translation never

Sectors/Track

changes the capacity of the drive; it

changes only the geometry to stay

Total Capacity

108 MB

Total Capacity

108 MB

within the BIOS limits.

Back to LBA Now let's watch how the advanced sector translation of LBA

provides support for hard drives greater than 504 MB. Let's use an old drive,

the Western Digital WD2160, a 2.1-GB hard drive, as an example. This drive is

no longer in production but its smaller CHS values make understanding LBA

easier. Table 8.3 lists its physical and logical geometries.

Note that, even with sector translation, the number of heads is greater than

the allowed 16! So here's where the magic of LBA comes in. The WD2160 is ca-

pable of LBA. Now assuming that the BIOS is also capable of LBA, here's

what happens. When the com-

puter boots up, the BIOS asks the

Table 8.3

Western Digital WD2160's Physical

drives if they can perform LBA. If

and Logical Geometries

they say yes, the BIOS and the

Physical

Logical

drive work together to change the

way they talk to each other. They

Cylinders

16,384

Cylinders

1024

can do this without conflicting

Heads

with the original AT BIOS com-

mands by taking advantage of un-

Sectors/Track

used commands to use up to 256

Total Capacity

2.1 GB

Total Capacity

2.1 GB

heads. LBA enables support for a

113

Chapter 8: Hard Drive Technologies

maximum of 1024 × 256 × 63 × 512 bytes = 8.4-GB hard drives. Back in 1990, 8.4

GB was hundreds of time larger than the drives used at the time. Don't worry,

later ATA standards will get the BIOS up to today's huge drives!

Not Just Hard Drives Anymore: ATAPI

ATA-2 added an extension to the ATA specification, called Advanced Tech-

With the introduction of

ATAPI, the ATA standards are

nology Attachment Packet Interface (ATAPI) , that enabled nonhard drive

often referred to as ATA/ATAPI

devices such as CD-ROM drives and tape backups to connect to the PC via

instead of just ATA.

the ATA controllers. ATAPI drives have the same 40-pin interface and mas-

ter/slave jumpers as ATA hard drives. Figure 8.20 shows an ATAPI

CD-RW drive attached to a motherboard. The key

difference between hard drives and every other

type of drive that attaches to the ATA controller is

in how the drives get BIOS support. Hard drives

get it through the system BIOS, whereas nonhard

drives require the operating system to load a

software driver.

More Drives with ATA-2

ATA-2 added support for a second controller,

raising the total number of supported drives from

two to four. Each of the two controllers is equal in

power and capability. Figure 8.21 is a close-up of

a typical motherboard, showing the primary con-

troller marked as IDE1 and the secondary marked

as IDE2.

Increased Speed

· Figure 8.20

ATAPI CD-RW drive attached to a motherboard via a

standard, 40-pin ribbon cable

ATA-2 defined two new PIO modes and a new

type of DMA called multi-word DMA that was a

substantial improvement over the old DMA.

· Figure 8.21

Primary and secondary controllers labeled on a motherboard

114

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Technically, multi-word DMA was still the old-style DMA, but it worked in

a much more efficient manner so it was much faster.

PIO mode 3: 11.1 MBps

■

PIO mode 4: 16.6 MBps

■

Multi-word DMA mode 0: 4.2 MBps

■

Multi-word DMA mode 1: 13.3 MBps

■

Multi-word DMA mode 2: 16.6 MBps

■

ATA-3

ATA-3 came on quickly after ATA-2 and added

one new feature called Self-Monitoring, Analy-

sis, and Reporting Technology (S.M.A.R.T. , one

of the few PC acronyms that requires the use of

periods after each letter). S.M.A.R.T. helps pre-

dict when a hard drive is going to fail by moni-

toring the hard drive's mechanical components.

S.M.A.R.T. is a great idea and is popular in

specialized server systems, but it's complex, im-

perfect, and hard to understand. As a result, only

a few utilities can read the S.M.A.R.T. data on

your hard drive. Your best sources are the hard

drive manufacturers. Every hard drive maker

has a free diagnostic tool (that usually works

only for their drives) that will do a S.M.A.R.T.

check along with other tests. Figure 8.22 shows

Western Digital's Data Lifeguard Tool in action.

Note that it says only whether the drive has

passed or not. Figure 8.23 shows some S.M.A.R.T.

· Figure 8.22

Data Lifeguard

information.

· Figure 8.23

S.M.A.R.T. information

115

Chapter 8: Hard Drive Technologies

Although you can see the actual S.M.A.R.T. data, it's generally useless or

indecipherable. It's best to trust the manufacturer's opinion and run the

software provided.

ATA-4

Anyone who has opened a big database file on a hard drive appreciates that

a faster hard drive is better. ATA-4 introduced a new DMA mode called

Ultra DMA that is now the primary way a hard drive communicates with a

PC. Ultra DMA uses DMA bus mastering to achieve far faster speeds than

was possible with PIO or old-style DMA. ATA-4 defined three Ultra DMA

modes:

Ultra DMA mode 0: 16.7 MBps

■

Ultra DMA mode 2, the most

popular of the ATA-4 DMA

Ultra DMA mode 1: 25.0 MBps

■

modes, is also called ATA/33.

Ultra DMA mode 2: 33.3 MBps

■

INT13 Extensions

Here's an interesting factoid for you: The original ATA-1 standard allowed

for hard drives up to 137 GB! It wasn't the ATA standard that caused the

504-MB size limit, it was the fact that the standard used the old AT BIOS and

the BIOS, not the ATA standard, could support only 504 MB. LBA was a

work-around that told the hard drive to lie to the BIOS to get it up to 8.4 GB.

But eventually hard drives started edging close to the LBA limit and some-

thing had to be done. The T13 folks said, "This isn't our problem! It's the an-

cient BIOS problem. You BIOS makers need to fix the BIOS!" And they did.

In 1994, Phoenix Technologies (the BIOS manufacturer) came up with a

new set of BIOS commands called Interrupt 13 (INT13) extensions . INT13

extensions broke the 8.4-GB barrier by completely ignoring the CHS values

and instead feeding the LBA a stream of addressable sectors. A system with

INT13 extensions can handle drives up to 137 GB. The entire PC industry

quickly adopted INT13 extensions and every system made since 20002001

supports INT13 extensions.

ATA-5

Ultra DMA was such a huge hit that the ATA folks adopted two faster Ultra

DMA modes with ATA-5:

Ultra DMA mode 4, the most

popular of the ATA-5 DMA

Ultra DMA mode 3: 44.4 MBps

■

modes, is also called ATA/66.

Ultra DMA mode 4: 66.6 MBps

■

Ultra DMA modes 4 and 5 ran so quickly that the ATA-5 standard de-

fined a new type of ribbon cable of handling the higher speeds. This 80-wire

cable still has 40 pins on the connectors, but it includes another 40 wires

in the cable that act as grounds to improve the cable's ability to handle

high-speed signals. The 80-wire cable, just like the 40-pin ribbon cable, has a

colored stripe down one side give you proper orientation for pin 1 on the

116

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

controller and the hard drive. Previous ver-

sions of ATA didn't define where the differ-

ent drives were plugged into the ribbon

cable, but ATA-5 defined exactly where the

controller, master, and slave drives con-

nected, even defining colors to identify

them. Take a look at the ATA/66 cable in

Figure 8.24. The connector on the left is col-

ored blue (which you could see if the photo

was in color!)--that connector must be the

one used to plug into the controller. The

connector in the middle is grey--that's for

the slave drive. The connector on the right is · Figure 8.24 ATA/66 cable

black--that's for the master drive. Any

ATA/66 controller connections are colored

blue to let you know it is an ATA/66

controller.

ATA/66 is backward compatible, so you may safely plug an earlier

drive into an ATA/66 cable and controller. If you plug an ATA/66 drive

into an older controller it will work⎯just not in ATA/66 mode. The only

risky action is to use an ATA/66 controller and hard drive with a non-ATA/

66 cable. Doing so will almost certainly cause nasty data losses!

ATA-6

Hard drive size exploded in the early 21st century and the seemingly

impossible-to-fill 137-GB limit created by INT13 extensions became a bar-

rier to fine computing more quickly than most people had anticipated.

When drives started hitting the 120-GB mark, the T13 committee adopted an

industry proposal pushed by Maxtor (a major hard drive maker) called Big

Drive that increased the limit to more than 144 petabytes (approximately

144,000,000 GB). T13 also thankfully gave the new standard a less-silly

name, calling it ATA/ATAPI-6 or simply ATA-6 . Big Drive was basically just

a 48-bit LBA, supplanting the older 24-bit addressing of LBA and INT13 ex-

tensions. Plus, the standard defined an enhanced block mode, enabling

drives to transfer up to 65,536 sectors in one chunk, up from the measly 256

sectors of lesser drive technologies.

ATA-6 also introduced Ultra DMA mode 5, kicking the data transfer rate

up to 100 MBps. Ultra DMA mode 5 is more commonly referred to as ATA/

100, which requires the same 80-wire connectors as ATA/66.

ATA-7

ATA-7 brought two new innovations to the ATA world⎯one evolutionary

and the other revolutionary. The evolutionary innovation came with the last

of the parallel ATA Ultra DMA modes; the revolutionary was a new form of

ATA called serial ATA (SATA).

117

Chapter 8: Hard Drive Technologies

ATA/133

ATA-7 introduced the fastest and probably least adopted of all the ATA

speeds, Ultra DMA mode 6 ( ATA/133 ). Even though it runs at a speed of

133 MBps, the fact that it came out with SATA kept many hard drive manu-

facturers away. ATA/133 uses the same cables as Ultra DMA 66 and 100.

While you won't find many ATA/133 hard drives, you will find plenty

of ATA/133 controllers. There's a trend in the industry to color the control-

ler connections on the hard drive red, although this is not part of the ATA-7

standard.

Serial ATA

The real story of ATA-7 is SATA. For all its longevity as the mass storage in-

terface of choice for the PC, parallel ATA has problems. First, the flat ribbon

cables impede airflow and can be a pain to insert properly. Second, the ca-

bles have a limited length, only 18 inches. Third, you can't hot-swap PATA

drives. You have to shut down completely before installing or replacing a

drive. Finally, the technology has simply reached the limits of what it can do

in terms of throughput.

Serial ATA addresses these issues. SATA creates a point-to-point connec-

tion between the SATA device--hard disk, CD-ROM, CD-RW, DVD-ROM,

DVD-RW, and so on--and the SATA controller. At a glance, SATA devices

look identical to standard PATA devices. Take a closer look at the cable and

power connectors, however, and you'll see significant differences (Figure 8.25).

Because SATA devices send data serially instead of in parallel, the

SATA interface needs far fewer physical wires--seven instead of the eighty

wires that is typical of PATA--resulting in much thinner cabling. This

might not seem significant, but the benefit is that thinner cabling means

better cable control and better airflow through the PC case, resulting in

better cooling.

Data port

Power port

Data cable

Power cable

· Figure 8.25

SATA hard drive cables and connectors

118

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Further, the maximum SATA device cable length is more than twice that

Tech Tip

of an IDE cable--about 40 inches (1 meter) instead of 18 inches. Again, this

might not seem like a big deal, unless you've struggled to connect a PATA

SATA Names

hard disk installed into the top bay of a full-tower case to an IDE connector

Number-savvy readers might

located all the way at the bottom of the motherboard!

have noticed a discrepancy be-

SATA does away with the entire master/slave concept. Each drive con-

tween the names and throughput

of the two SATA drives. After all,

nects to one port, so no more daisy-chaining drives. Further, there's no maxi-

mum number of drives⎯many motherboards are now available that support

1.5 Gb per second throughput

translates to 192 MB per second,

up to eight SATA drives. Want more? Snap in a SATA host card and load

a lot higher than the advertised

'em up!

speed of a "mere" 150 MBps. The

The big news, however, is in data throughput. As the name implies,

same is true of the 3Gb/300 MBps

SATA devices transfer data in serial bursts instead of parallel, as PATA de-

drives. The encoding scheme used

vices do. Typically, you might not think of serial devices as being faster than

on SATA drives takes about

parallel, but in this case, that's exactly the case. A SATA device's single

20 percent of the overhead for the

stream of data moves much faster than the multiple streams of data coming

drive, leaving 80 percent for pure

from a parallel IDE device--theoretically up to 30 times faster! SATA drives

bandwidth. The 3Gb drive created

come in two common varieties, the 1.5Gb and the 3Gb, that have a maxi-

all kinds of problems, because the

mum throughput of 150 MBps and 300 MBps, respectively.

name of the committee working

on the specifications was called

SATA is backward compatible with current PATA standards and en-

the SATA II committee, and mar-

ables you to install a parallel ATA device, including a hard drive, optical

keters picked up on the SATA II

drive, and other devices, to a serial ATA controller by using a SATA bridge .

name. As a result, you'll find

A SATA bridge manifests as a tiny card that you plug directly into the

many brands called SATA II

40-pin connector on a PATA drive. As you can see in Figure 8.26, the con-

rather than 3Gb. The SATA com-

troller chip on the bridge requires separate power; you plug a Molex con-

mittee now goes by the name

nector into the PATA drive as normal. When you boot the system, the PATA

SATA-IO.

drive shows up to the system as a SATA drive.

SATA's ease of use has made it the choice for desktop system storage,

and its success is already showing in the fact that more than 90 percent of all

hard drives sold today are SATA drives.

· Figure 8.26

SATA bridge

119

Chapter 8: Hard Drive Technologies

eSATA

External SATA (eSATA) extends

the SATA bus to external devices,

as the name would imply. The

eSATA drives use similar connec-

tors to internal SATA, but they're

keyed differently so you can't mis-

take one for the other. Figure 8.27

shows eSATA connectors on the

back of a motherboard. eSATA uses

shielded cable lengths up to 2 me-

ters outside the PC and is hot

pluggable. The beauty of eSATA is

that it extends the SATA bus at full

speed, so you're not limited to the

meager 50 or 60 MBps of FireWire

or USB.

· Figure 8.27

eSATA connectors (center; that's a FireWire port on the left)

SCSI: Still Around

■

Many specialized server machines and enthusiasts' systems use the small

computer system interface (SCSI) technologies for various pieces of core

hardware and peripherals, from hard drives to printers to high-end tape

The CompTIA A+ 220-604

backup machines. SCSI is different from ATA in that SCSI devices connect

exam tests you on SCSI and

together in a string of devices called a chain. Each device in the chain gets a

RAID, topics essential to server

SCSI ID to distinguish it from other devices on the chain. Last, the ends of a

environments.

SCSI chain must be terminated. Let's dive into SCSI now, and see how SCSI

chains, SCSI IDs, and termination all work.

SCSI is an old technology dating from the late 1970s, but it has been con-

tinually updated. SCSI is faster than ATA

(though the gap is closing fast), and until

SATA arrived SCSI was the only good

choice for anyone using RAID (see the

"RAID" section a little later). SCSI is ar-

guably fading away, but it still deserves

some mention.

SCSI Chains

SCSI manifests itself through a SCSI

chain , a series of SCSI devices working

together through a host adapter. The host

adapter provides the interface between

the SCSI chain and the PC. Figure 8.28

shows a typical PCI host adapter. Many

techs refer to the host adapter as the SCSI

controller, so you should be comfortable

· Figure 8.28

with both terms.

SCSI host adapter

120

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

All SCSI devices can be divided

into two groups: internal and external.

Internal SCSI devices are attached in-

side the PC and connect to the host

adapter through the latter's internal

connector. Figure 8.29 shows an inter-

nal SCSI device, in this case a CD-ROM

drive. External devices hook to the ex-

ternal connector of the host adapter.

Figure 8.30 is an example of an external

SCSI device.

Internal SCSI devices connect to

the host adapter with a 68-pin ribbon

cable (Figure 8.31). This flat, flexible ca-

ble functions precisely like a PATA

cable. Many external devices connect to

the host adapter with a 50-pin high

density (HD) connector. Figure 8.32

shows a host adapter external port.

Higher end SCSI devices use a 68-pin

high density (HD) connector.

· Figure 8.29 Internal SCSI CD-ROM

Multiple internal devices can be

connected together simply by using a

cable with enough connectors. Figure 8.33, for example, shows a cable that

can take up to four SCSI devices, including the host adapter.

Assuming the SCSI host adapter has a standard external port (some con-

trollers don't have external connections at all), plugging in an external SCSI

device is as simple as running a cable from device to controller. The external

SCSI connectors are D-shaped, so you can't plug them in backward. As an

added bonus, some external SCSI devices have two ports, one to connect to

the host adapter and a second to connect to another SCSI device. The process

of connecting a device directly to another device is called daisy-chaining.

· Figure 8.30

· Figure 8.31

Back of external SCSI device

Typical 68-pin ribbon cable

121

Chapter 8: Hard Drive Technologies

You can daisy-chain up to 15 devices to one host adapter. SCSI

chains can be internal, external, or both (see Figure 8.34).

SCSI IDs

If you're going to connect a number of devices on the same

SCSI chain, you must provide some way for the host adapter to

tell one device from another. To differentiate devices, SCSI

uses a unique identifier called the SCSI ID . The SCSI ID num-

ber can range from 0 to 15. SCSI IDs are similar to many other

· Figure 8.32 50-pin HD port on SCSI host adapter

PC hardware settings in that a SCSI device can theoretically

have any SCSI ID, as long as that ID is not already taken by an-

other device connected to the same host adapter.

Some conventions should be followed when setting SCSI IDs. Typically,

Old SCSI equipment allowed

most people set the host adapter to 7 or 15, but you can change this setting.

SCSI IDs from 0 to 7 only.

Note that there is no order for the use of SCSI IDs. It does not matter which

device gets which number, and you

can skip numbers. Restrictions on

IDs apply only within a single chain.

Two devices can have the same ID, in

other words, as long as they are on

different chains (Figure 8.35).

Every SCSI device has some

method of setting its SCSI ID. The

trick is to figure out how as you're

holding the device in your hand. A

SCSI device may use jumpers, dip

switches, or even tiny dials; every

new SCSI device is a new adventure

as you try to determine how to set its

SCSI ID.

· Figure 8.33

Internal SCSI chain with two devices

· Figure 8.34

Internal and external devices on one SCSI chain

122

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Termination

Whenever you send a signal down a wire, some of

that signal reflects back up the wire, creating an

echo and causing electronic chaos. SCSI chains use

termination to prevent this problem. Termination

simply means putting something on the ends of the

wire to prevent this echo. Terminators are usually

pull-down resistors and can manifest themselves in

many different ways. Most of the devices within a

PC have the appropriate termination built in. On

other devices, including SCSI chains and some net-

work cables, you have to set termination during

installation.

The rule with SCSI is that you must terminate · Figure 8.35 IDs don't conflict between separate SCSI chains.

only the ends of the SCSI chain. You have to termi-

nate the ends of the cable, which usually means that

you need to terminate the two devices at the ends of the cable. Do not termi-

nate devices that are not on the ends of the cable. Figure 8.36 shows some ex-

amples of where to terminate SCSI devices.

Because any SCSI device might be on the end of a chain, most manufac-

turers build SCSI devices that can self-terminate. Some devices will detect

that they are on the end of the SCSI chain and will automatically terminate

themselves. Most devices, however, require that you set a jumper or switch to

enable termination (Figure 8.37).

· Figure 8.36

· Figure 8.37

Location of the terminated devices

Setting termination

123

Chapter 8: Hard Drive Technologies

Protecting Data with RAID

■

Ask experienced techs, "What is the most expensive part of a PC?" and

they'll all answer in the same way: "It's the data." You can replace any sin-

gle part of your PC for a few hundred dollars at most, but if you lose critical

data--well, let's just say I know of two small companies that went out of

business just because they lost a hard drive full of data.

Data is king; data is your PC's raison d'être. Losing data is a bad thing, so

you need some method to prevent data loss. Now, of course, you can do

backups, but if a hard drive dies, you have to shut down the computer, rein-

stall a new hard drive, reinstall the operating system, and then restore the

backup. There's nothing wrong with this as long as you can afford the time

and cost of shutting down the system.

A better solution, though, would save your data if a hard drive died and

enable you to continue working throughout the process. This is possible if

you stop relying on a single hard drive and instead use two or more drives

to store your data. Sounds good, but how do you do this? Well, first of all,

you could install some fancy hard drive controller that reads and writes

data to two hard drives simultaneously (Figure 8.38). The data on each drive

would always be identical. One drive would be the primary drive and the

other drive, called the mirror drive, would not be used unless the primary

drive failed. This process of reading and writing data at the same time to

two drives is called disk mirroring .

If you really want to make data safe, you can use two separate control-

lers for each drive. With two drives, each on a separate controller, the

system will continue to operate,

even if the primary drive's control-

ler stops working. This super-drive

mirroring technique is called disk

duplexing (Figure 8.39). Disk du-

plexing is also much faster than

disk mirroring because one control-

ler does not write each piece of data

twice.

Even though duplexing is faster

than mirroring, they both are

slower than the classic one drive,

one controller setup. You can use

multiple drives to increase your

hard drive access speed. Disk strip-

ing (without parity) means spread-

ing the data among multiple (at

least two) drives. Disk striping by

itself provides no redundancy. If

you save a small Microsoft Word

file, for example, the file is split into

multiple pieces; half of the pieces go

on one drive and half on the other

(Figure 8.40).

· Figure 8.38

Mirrored drives

124

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

· Figure 8.39

Duplexing drives

· Figure 8.40

Disk striping

125

Chapter 8: Hard Drive Technologies

The one and only advantage of disk striping is speed--it is a fast way to

read and write to hard drives. But if either drive fails, all data is lost. Disk

striping is not something you should do--unless you're willing to increase

the risk of losing data to increase the speed at which your hard drives save

and restore data.

Disk striping with parity , in contrast, protects data by adding extra infor-

mation, called parity data, that can be used to rebuild data should one of the

drives fail. Disk striping with parity requires at least three drives, but it is

common to use more than three. Disk striping with parity combines the best

of disk mirroring and plain disk striping. It protects data and is quite fast.

The majority of network servers use a type of disk striping with parity.

RAID

A couple of sharp guys in Berkeley back in the 1980s organized the many

An array in the context of

techniques for using multiple drives for data protection and increasing

RAID refers to a collection of two

speeds as the redundant array of independent (or inexpensive) disks (RAID) .

or more hard drives.

They outlined seven levels of RAID, numbered 0 through 6.

RAID 0--Disk Striping Disk striping requires at least two drives.

■

It does not provide redundancy to data. If any one drive fails, all

data is lost.

RAID 1--Disk Mirroring/Duplexing RAID 1 arrays require at

■

least two hard drives, although they also work with any even

number of drives. RAID 1 is the ultimate in safety, but you lose

storage space since the data is duplicated--you need two 100-GB

drives to store 100 GB of data.

RAID 2--Disk Striping with Multiple Parity Drives RAID 2 was

■

a weird RAID idea that never saw practical use. Unused, ignore it.

RAID 3 and 4--Disk Striping with Dedicated Parity RAID 3 and

■

4 combined dedicated data drives with dedicated parity drives. The

differences between the two are trivial. Unlike RAID 2, these

versions did see some use in the real world but were quickly

replaced by RAID 5.

RAID 5--Disk Striping with Distributed Parity Instead of

■

dedicated data and parity drives, RAID 5 distributes data and parity

information evenly across all drives. This is the fastest way to

Tech Tip

provide data redundancy. RAID 5 is by far the most common RAID

implementation and requires at least three drives. RAID 5 arrays

RAID Lingo

effectively use one drive's worth of space for parity. If, for example,

No tech worth her salt says

you have three 200-GB drives, your total storage capacity is 400 GB.

things like, "We're implementing

If you have four 200-GB drives, your total capacity is 600 GB.

disk striping with parity." Use

the RAID level. Say, "We're im-

RAID 6--Disk Striping with Extra Parity If you lose a hard drive

■

plementing RAID 5." It's more

in a RAID 5 array, your data is at great risk until you replace the bad

accurate and very impressive to

hard drive and rebuild the array. RAID 6 is RAID 5 with extra parity

the folks in the accounting

information. RAID 6 needs at least five drives, but in exchange you

department!

can lose up to two drives at the same time. RAID 6 is gaining in

popularity for those willing to use larger arrays.

126

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

After these first RAID levels were defined, some manufacturers came up

with ways to combine different RAIDs. For example, what if you took two

You'll hear the term nested

pairs of striped drives and mirrored the pairs? You would get what is called

RAID used for multiple RAID

solutions. The terms are

RAID 0+1. Or what (read this carefully now) if you took two pairs of mir-

synonymous.

rored drives and striped the pairs? You then get what we call RAID 1+0 or

what is often called RAID 10. Combinations of different types of single

RAID are called multiple RAID solutions. Multiple RAID solutions, while

enjoying some support in the real world, are quite rare when compared to

There is actually a term for a

single RAID solutions RAID 0, 1, and 5.

storage system composed of

multiple independent disks,

rather than disks organized us-

Implementing RAID

ing RAID: JBOD, which stands

for Just a Bunch of Disks (or

RAID levels describe different methods of providing data redundancy or

Drives).

enhancing the speed of data throughput to and from groups of hard drives.

They do not say how to implement these methods. Literally thousands of dif-

ferent methods can be used to set up RAID. The method used depends

largely on the desired level of RAID, the operating system used, and the

thickness of your wallet.

The obvious starting place for RAID is to connect at least two hard

drives in some fashion to create a RAID array. For many years, if you

wanted to do RAID beyond RAID 0 and RAID 1, the only technology you

could use was good-old SCSI. SCSI's chaining of multiple devices to a single

controller made it a natural for RAID. SCSI drives make superb RAID ar-

rays, but the high cost of SCSI drives and RAID-capable host adapters kept

RAID away from all but the most critical systems--usually big file servers.

In the last few years, substantial leaps in ATA technology have made

ATA a viable alternative to SCSI drive technology for RAID arrays. Special-

ized ATA RAID controller cards support ATA RAID arrays of up to 15

drives--plenty to support even the most complex RAID needs. In addition,

the inherent hot-swap capabilities of serial ATA have virtually guaranteed

that serial ATA will quickly take over the lower end of the RAID business.

Try This!

Managing Heat with Multiple Drives

Adding three or more fast hard drives into a cramped PC case can be a

recipe for disaster to the unwary tech. All those disks spinning con-

stantly create a phenomenal amount of heat. Heat kills PCs! You've got

to manage the heat inside a RAID-enabled system or risk losing your

data, drives, and basic system stability. The easiest way to do this is to

add fans, so Try This!

Open up your PC case and look for built-in places to mount fans.

How many case fans do you have installed now? What size are they?

What sizes can you use? (Most cases use 80 mm fans, but 60 and 120 mm

fans are common as well.) Jot down the particulars of your system and

take a trip to the local PC store to check out the fans.

Before you get all fan-happy and grab the biggest and baddest fans

to throw in your case, don't forget to think about the added noise level.

Try to get a compromise between keeping your case cool enough and

not causing early deafness!

127

Chapter 8: Hard Drive Technologies

Personally, I think the price and performance of serial ATA mean SCSI's

days are numbered.

Once you have a number of hard drives, the next question is whether to

use hardware or software to control the array. Let's look at both options.

Hardware versus Software

You can use Disk Manage-

ment in Windows 2000 and Win-

All RAID implementations break down into either hardware or software meth-

dows XP Professional to create

ods. Software is often used when price takes priority over performance. Hard-

RAID 1 and RAID 5 arrays, but

ware is used when you need speed along with data redundancy. Software

you can use Disk Management

only remotely on a server ver-

RAID does not require special controllers--you can use the regular ATA con-

sion of Windows (2000 Server or

trollers or SCSI host adapters to make a software RAID array. But you do need

Server 2003). In other words, the

"smart" software. The most common software implementation of RAID is the

capability is there, but Microsoft

built-in RAID software that comes with Windows 2000 Server and Windows

has limited the OS. If you want

Server 2003. The Disk Management program in these Windows Server ver-

to use software RAID in Win-

dows 2000 or XP (Home or Pro-

sions can configure drives for RAID 0, 1, or 5, and it works with ATA or SCSI

fessional), you need to use a

(Figure 8.41). Disk Management in Windows 2000 Professional and Windows

third-party tool to set it up.

XP Professional, in contrast, can only do RAID 0.

Windows Disk Management is not the only software RAID game in

town. A number of third-party software programs can be used with Win-

dows or other operating systems.

Software RAID means the operating system is in charge of all RAID

functions. It works for small RAID solutions but tends to overwork your op-

erating system easily, creating slowdowns. When you really need to keep

going, when you need RAID that doesn't even let the users know a problem

has occurred, hardware RAID is the answer.

Hardware RAID centers around an intelligent controller--either a SCSI

host adapter or an ATA controller that handles all of the RAID functions

(Figure 8.42). Unlike a regular ATA controller or SCSI host adapter, these

controllers have chips that know how to "talk RAID."

· Figure 8.41

Disk Management in Windows Server 2003

128

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

· Figure 8.42

Serial ATA RAID controller

Most RAID setups in the real world are hardware-based. Almost all of

the many hardware RAID solutions provide hot-swapping--the ability to re-

place a bad drive without disturbing the operating system. Hot-swapping is

common in hardware RAID.

Hardware-based RAID is invisible to the operating system and is config-

ured in several ways, depending on the specific chips involved. Most RAID

systems have a special configuration utility in Flash ROM that you access af-

ter CMOS but before the OS loads. Figure 8.43 shows a typical firmware pro-

gram used to configure a hardware RAID solution.

Personal RAID

Due to drastic reductions in the cost of ATA RAID controller chips, in

the last few years we've seen an explosion of ATA-based hardware RAID

· Figure 8.43

RAID configuration utility

129

Chapter 8: Hard Drive Technologies

solutions built into mainstream motherboards. While this "ATA RAID on

the motherboard" began with parallel ATA, the introduction of serial ATA

RAID controllers aren't just

made motherboards with built-in RAID extremely common.

for internal drives; some models

can handle multiple eSATA

These personal RAID motherboards might be quite common, but

drives configured using any of

they're not used too terribly often given that these RAID solutions usually

the different RAID levels. If

provide only RAID 0 or RAID 1. If you want to use RAID, spend a few extra

you're feeling lucky, you can cre-

dollars and buy a RAID 5capable controller.

ate a RAID array using both in-

ternal and external SATA drives.

The Future Is RAID

RAID has been with us for about 20 years, but until only recently it was the

domain of big systems and deep pockets. During those 20 years, however, a

number of factors have come together to make RAID a reality for both big

servers and common desktop systems. Imagine a world where dirt-cheap

RAID on every computer means no one ever again losing critical data. I get

goose bumps just thinking about it!

Connecting Drives

■

Installing a drive is a fairly simple process if you take the time to make sure

you've got the right drive for your system, configure the drive properly,

and do a few quick tests to see if it's running properly. Since PATA, SATA,

and SCSI have different cabling requirements, we'll look at each of these

separately.

Choosing Your Drive

First, decide where you're going to put the drive. Look for an open ATA

connection. Is it PATA or SATA? Is it a dedicated RAID controller? Many

motherboards with built-in RAID controllers have a CMOS setting that en-

ables you to turn the RAID on or off (Figure 8.44). Do you have the right con-

troller for a SCSI drive?

· Figure 8.44

Settings for RAID in CMOS

130

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Second, make sure you have room for the drive in the case. Where will

you place it? Do you have a spare power connector? Will the data and power

cables reach the drive? A quick test fit is always a good idea.

Don't worry about PIO Modes and DMA⎯a new drive will support

anything your controller wants to do.

Jumpers and Cabling

on PATA Drives

If you have only one hard drive, set the drive's

jumpers to master or standalone. If you have two

drives, set one to master and the other to slave. See

Figure 8.45 for a close-up of a PATA hard drive

showing the jumpers.

At first glance, you might notice that the jump-

ers aren't actually labeled master and slave. So how

do you know how to set them properly? The easi-

est way is to read the front of the drive; most

drives have a diagram on the housing that explains

how to set the jumpers properly. Figure 8.46 shows

the front of one of these drives, so you can see how

to set the drive to master or slave.

Hard drives may have other jumpers that may

or may not concern you during installation. One · Figure 8.45 Master/slave jumpers on a hard drive

common set of jumpers is used for diagnostics at

the manufacturing plant or for special settings in

other kinds of devices that use hard drives. Ignore

them. They have no bearing in the PC world. Second, many drives provide a

third setting, which is used if only one drive connects to a controller. Often,

Most of the high speed ATA/

66/100/133 cables support cable

master and single drive are the same setting on the hard drive, although

select--try one and see!

some hard drives require separate settings. Note that the name for the single

drive setting varies among manufacturers. Some use Single; others use 1

Drive or Standalone.

Many current PATA hard drives use a jumper setting called cable select,

rather than master or slave. As the name implies, the position on the cable

determines which drive will be master or slave: master on the end, slave in

the middle. For cable select to work properly with two drives, both

drives must be set as cable select and the cable itself must be a spe-

cial cable-select cable. If you see a ribbon cable with a pinhole

through one wire, watch out! That's a cable-select cable.

If you don't see a label on the drive that tells you how to set the

jumpers, you have several options. First, look for the drive maker's

Web site. Every drive manufacturer lists its drive jumper settings on

the Web, although it can take a while to find the information you

want. Second, try phoning the hard drive maker directly. Unlike

many other PC parts manufacturers, hard drive producers tend to

stay in business for a long period of time and offer great technical

support.

Hard drive cables have a colored stripe that corresponds to the · Figure 8.46 Drive label showing master/slave

number-one pin--called pin 1--on the connector. You need to make

settings

131

Chapter 8: Hard Drive Technologies

certain that pin 1 on the controller

Cross Check

is on the same wire as pin 1 on the

hard drive. Failing to plug in the

Molex Connectors

drive properly will also prevent

Hard drives and other internal devices use Molex connectors for power.

the PC from recognizing the drive.

Check your memory from the Essentials course. What voltages go

If you incorrectly set the master/

through the four wires on a Molex connector? What should you note

slave jumpers or cable to the hard

about the connector and inserting it into the corresponding socket on

drives, you won't break anything;

the drive?

it just won't work.

Finally, you need to plug a

Molex connector from the power supply into the drive. All modern PATA

drives use a Molex connector.

Cabling SATA Drives

Installing SATA hard drives is even easier than installing IDE devices due to

the fact that there's no master, slave, or cable select configuration to mess

with. In fact, there are no jumper settings to worry about at

all, as SATA supports only a single device per controller

channel. Simply connect the power and plug in the control-

ler cable as shown in Figure 8.47--the OS automatically de-

tects the drive and it's ready to go! The keying on SATA

controller and power cables makes it impossible to install

either incorrectly.

The biggest problem with SATA drives is that many

motherboards come with four or more. Sure, the cabling is

easy enough, but what do you do when it comes time to

start the computer and the system is trying to find the right

hard drive to boot up! That's where CMOS comes into

play.

Connecting SCSI Drives

· Figure 8.47

Properly connected SATA cable

Connecting SCSI drives requires three things. You must

use a controller that works with your drive. You need to set

unique SCSI IDs on the controller and the drive. Finally,

you need to connect the ribbon cable and power connections properly. With

SCSI, you need to attach the data cable correctly. You can reverse a PATA

cable, for example, and nothing happens except the drive doesn't work. If

you reverse a SCSI cable, however, you can seriously damage the drive. Just

as with PATA cables, pin 1 on the SCSI data cable must go to pin 1 on both

the drive and the host adapter.

BIOS Support: Configuring CMOS

■

and Installing Drivers

Every device in your PC needs BIOS support, and the hard drive controllers

are no exception. Motherboards provide support for the ATA hard drive

132

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

controllers via the system BIOS, but they require configuration in CMOS for

the specific hard drives attached. SCSI drives require software drivers or

firmware on the host adapter.

In the old days, you had to fire up CMOS and manually enter CHS infor-

mation whenever you installed a new ATA drive to ensure the system saw

the drive. Today, this process still takes place, but it's much more auto-

mated. Still, there's plenty to do in CMOS when you install a new hard

drive.

CMOS settings for hard drives vary a lot among motherboards. The fol-

lowing information provides a generic look at the most common settings,

but you'll need to look at your specific motherboard manual to understand

all the options available.

Configuring Controllers

As a first step in configuring controllers, make certain they're enabled. It's

easy to turn off controllers in CMOS, and many motherboards turn off sec-

ondary ATA controllers by default. Scan through your CMOS settings to lo-

cate the controller on/off options (see Figure 8.48 for typical settings). This

is also the time to check whether your onboard RAID controllers work in

both RAID and non-RAID settings.

Autodetection

If the controllers are enabled and the drive is properly connected, the drive

should appear in CMOS through a process called autodetection. Autodetection

is a powerful and handy feature, but it seems every CMOS has a different

way to manifest it, and how it is manifested may affect how your computer

decides which hard drive to try to boot when you start your PC.

One of your hard drives stores the operating system needed when you

boot your computer, and your system needs a way to know where to look

for this operating system. The traditional BIOS supported a maximum of

only four ATA drives on two controllers, called the primary controller and the

secondary controller. The BIOS

looked for the master drive on the

primary controller when the system

booted up. If you used only one

controller, you used the primary

controller. The secondary controller

was used for CD-ROMs, DVDs, or

other non-bootable drives.

Older CMOS made this clear

and easy, as shown in Figure 8.49.

When you booted up, the CMOS

would query the drives through

autodetection, and whatever drives

the CMOS saw would show up

here. Some even older CMOS had a

special

option

called

Autodetect that you had to run to · Figure 8.48 Typical controller settings in CMOS

133

Chapter 8: Hard Drive Technologies

see the drives in this screen. There

are places for up to four de-

vices⎯notice not all of them actu-

ally have a device.

The autodetection screen indi-

cated that you installed a PATA

drive correctly. If you installed a

hard drive on the primary controller

as master, but messed up the jumper

and set it to slave, it would show up

in the autodetection screen as the

slave. If you had two drives and set

them both to master, one drive or the

other (or sometimes both) didn't ap-

pear, telling you that something was

messed up in the physical installa-

· Figure 8.49

Old standard CMOS settings

tion. If you forgot to plug in the rib-

bon cable or the power, the drives

wouldn't autodetect.

SATA messed up the autodetection happiness. There's no such thing as

master, slave, or even primary and secondary controller in the SATA world.

To get around this, motherboards with PATA and SATA today use a num-

bering system⎯and every motherboard uses its own numbering system!

One common numbering method uses the term channels for each controller.

The first boot device is channel 1, the second is channel 2, and so on. PATA

channels may have a master and a slave, but a SATA channel has only a

master, as SATA controllers support only one drive. So instead of names of

drives, you see numbers. Take a look at Figure 8.50.

Whew! Lots of hard drives! This motherboard supports the traditional

four PATA drives, but it also supports four SATA drives. Each controller is

assigned a number⎯note that channel 1 and channel 2 have master/slave

settings, and that's how you know channel 1 and 2 are the PATA drives.

Channels 3 through 6 are SATA, even though the listing says master.

(SATA's still somewhat new, and a

CMOS using incorrect terms like

master is common.)

Boot Order

If you want your computer to run,

it's going to need an operating sys-

tem to boot. While the PCs of our

forefathers (those of the 1980s and

early 1990s) absolutely required

you to put the operating system on

the primary master, most BIOS

makers by 1995 enabled you to put

the OS on any of the four drives and

then tell the system through CMOS

which hard drive to boot. Addition-

· Figure 8.50

New standard CMOS features

ally, you may need to boot from

134

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

a floppy, CD-ROM, or even a thumb drive at times. CMOS takes

care of this by enabling you to set a boot order.

Figure 8.51 shows a typical boot order screen. It has a first, sec-

ond, and third boot option. Many users like to boot first from floppy

or CD-ROM, and then from a hard drive. This enables them to put in

· Figure 8.51 Boot order

a bootable floppy or CD-ROM if they're having problems with the

system. Of course, you can set it to boot first from your hard drive

and then go into CMOS and change it when you need to⎯it's your

choice.

Most modern CMOS lump the

Try This!

hard drive boot order onto a sec-

ond screen. This screen works like

Working with CMOS

an autodetect in that it shows only

One of the best ways to get your mind around the different drive stan-

actual hard drives attached. This

dards and capabilities is to run benchmarking software on the hard

beats the heck out of guessing!

drive to get a baseline of its capabilities. Then, change CMOS settings to

alter the performance of the drive and run the diagnostics again. Try

Device Drivers

this!

Devices that do not get BIOS via the

Get a reliable hard drive benchmarking program. I recommend

system BIOS routines naturally re-

HD Tach (www.simplisoftware.com) as reliable and rugged.

quire some other source for BIOS.

Run the software, and record the scores.

For ATAPI devices and many

Change some or all of the following CMOS settings, and then

SATA controllers, the source of

run the benchmarking utility again: PIO mode, DMA mode,

choice is software device drivers,

Block mode.

but both technologies have a couple

What were the effects of changing settings?

of quirks you should know about.

ATAPI Devices and BIOS

ATAPI drives plug into an ATA controller on the motherboard and follow

the same conventions on cabling and jumpers used by PATA hard drives. In

fact, all current CMOS setup utilities seem to autodetect CD-media ATAPI

drives. If you go into CMOS after

installing a CD-ROM drive as mas-

ter on the secondary IDE controller,

for example, the drive will show up

just fine, as in Figure 8.52.

The reporting of installed

CD-media drives in CMOS serves

two purposes. First, it tells the tech-

nician that he or she has good con-

nectivity on the ATAPI drive.

Second, it shows that you have the

option to boot to CD-media, such as

a Windows XP disc. What it doesn't

do, however, is provide true BIOS

support for that drive! That has to

come with a driver loaded at

· Figure 8.52 CMOS screen showing a CD-ROM drive detected

boot-up.

135

Chapter 8: Hard Drive Technologies

Troubleshooting Hard Drive

■

Installation

The best friend a tech has when it comes to troubleshooting hard drive in-

stallation is the autodetection feature of the CMOS setup utility. When a

The CompTIA A+ 220-603

drive doesn't work, the biggest question, especially during installation, is,

exam is interested in trouble-

"Did I plug it in correctly?" With autodetection, the answer is simple; if it

shooting storage devices, but

doesn't see the drives, something is wrong with the hardware configura-

more from the software side

tion. Either a device has physically failed or, more likely, you didn't give the

than the physical installation

hard drive power, plugged a cable in backwards, or messed up some other

angle.

connectivity issue.

It takes four things to get a drive installed and recognized by the system:

jumpers (PATA only), data cable, power, and CMOS setup recognizing the

drive. If any of these steps is missed or messed up, you have a drive that

simply doesn't exist according to the PC! To troubleshoot hard drives, sim-

ply work your way through each step to figure out what went wrong.

First, set the drive to master, slave, standalone, or cable select, depend-

ing on where you decide to install the drive. If a drive is alone on the cable,

set it to master or standalone. With two drives, one must be master and the

other slave. Alternatively, you can set both drives to cable select and use a

cable-select cable.

Second, the data cable must be connected to both the drive and control-

ler, Pin 1 to Pin 1. Reversing the data cable at one end is remarkably easy to

do, especially with the rounded cables. They obviously don't have a big red

stripe down the side to indicate the location of Pin 1! If you can't autodetect

the drive, check the cabling.

Third, be sure to give the hard drive power. Most hard drives use a stan-

dard Molex connector. If you don't hear the whirring of the drive, make cer-

tain you plugged in a Molex from the power supply, rather than from

another source such as an otherwise disconnected fan. You'd be surprised

how often I've seen that!

Fourth, you need to provide BIOS for the controller and the drive. This

can get tricky as the typical CMOS setup program has a lot of hard drive op-

tions. Plus, you have an added level of confusion with RAID settings and

non-integrated controllers that require software drivers.

Once you've checked the physical connections, run through these issues

in CMOS. Is the controller enabled? Is the storage technology--LBA, INT13,

ATA/ATAPI-6--properly set up? Similarly, can the motherboard support

the type of drive you're installing? If not, you have a couple of options. You

can flash the BIOS with an upgraded BIOS from the manufacturer or you

can get a hard drive controller that goes into an expansion slot.

Finally, make certain with non-integrated hard drive controllers, such as

those that come with many SATA drives, that you've installed the proper

drivers for the controller. Driver issues can crop up with new, very large

drives, and with changes in technology. Always check the manufacturer's

Web site for new drivers.

136

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Beyond A+

Spindle (or Rotational) Speed

Hard drives run at a set spindle speed, measured in revolutions per minute

(RPM). Older drives run at the long-standard speed of 3600 RPM, but new

drives are hitting 15,000 RPM! The faster the spindle speed, the faster the

controller can store and retrieve data. Here are the common speeds: 4500,

5400, 7200, and 10,000 RPM.

Faster drives mean better system performance, but they can also cause

the computer to overheat. This is especially true in tight cases, such as

minitowers, and in cases containing many drives. Two 4500 RPM drives

might run forever, snugly tucked together in your old case. But slap a hot

new 10,000 RPM drive in that same case and watch your system start crash-

ing right and left!

You can deal with these hotrod drives by adding drive bay fans between

the drives or migrating to a more spacious case. Most enthusiasts end up do-

ing both. Drive bay fans sit at the front of a bay and blow air across the drive.

They range in price from $10 to $100 (U.S.) and can lower the temperature of

your drives dramatically. Figure 8.53 shows a picture of a double-fan drive

bay cooler.

Air flow in a case can make or break your system

stability, especially when you add new drives that

increase the ambient temperature. Hot systems get

flaky and lock up at odd moments. Many things can

impede the air flow--jumbled up ribbon cables,

drives squished together in a tiny case, fans clogged

by dust or animal hair, and so on.

Technicians need to be aware of the dangers

when adding a new hard drive to an older system.

Get into the habit of tying off ribbon cables, adding

front fans to cases when systems lock up intermit-

tently, and making sure the fan(s) run well. Finally,

if a client wants a new drive and his system is a tiny

minitower with only the power supply fan to cool it

off, be gentle, but definitely steer him to one of the

slower drives!

· Figure 8.53 Bay fans

Hybrid Hard Drives

Windows Vista supports hybrid hard drives (HHDs), drives that combine

flash memory and spinning platters to provide fast and reliable storage.

Samsung has drives with 128-MB and 256-MB flash cache, for example, that

shave boot times in half and, because the platters don't have to spin all the

time, add 2030 minutes more of battery life for portable computers. Add-

ing that much more run time with only a tiny price premium and no extra

weight is the Holy Grail of portable computing!

137

Chapter 8: Hard Drive Technologies

Chapter 8 Review

■ Chapter Summary

After reading this chapter and completing the

geometric values, have no relevance in today's

exercises, you should understand the following about

PCs, but most CMOS utilities still support them.

hard drive technologies.

ATA--The King

How Hard Drives Work

Today's hard drives have either an ATA interface

■

or a SCSI interface. ATA drives may be parallel

Hard drives contain aluminum platters coated

■

ATA (PATA) or the newer serial ATA (SATA).

with a magnetic medium and read/write heads

A specification of the American National

that float on a cushion of air. Hard drives store

Standards Institute (ANSI), the AT Attachment

data in a tiny magnetic field called a flux that

(ATA) interface (commonly but incorrectly referred

defines a zero or a one. The switching back and

to as IDE) used a 40-pin ribbon cable, had a built-in

forth of the field is called flux reversal. The

controller, and did not require a low-level format.

incredible storage capacity of today's drives is due

By 1995, EIDE was the dominant interface. Its

to Partial Response Maximum Likelihood (PRML)

features include higher capacities, support for

encoding that includes intelligent circuitry to

nonhard drive storage devices, a four device

analyze each flux reversal.

maximum, and improved throughput. The terms

Two different technologies have been used to move

■

ATA, IDE, and EIDE are used interchangeably to

read/write heads across the platters: stepper

describe all PATA devices.

motors and voice coil. Very susceptible to physical

deterioration and temperature changes, the now

PATA drives use a 40-pin plug and a controller that

■

obsolete stepper motors moved the actuator arm in

connects them to the external data bus. Although

fixed increments or steps, often resulting in data

the real controller is built into the hard drive itself,

transfer errors or the inability to access data on a

the 40-pin connector on the motherboard is

cold drive. The heads had to be parked to a

called the controller. Most modern motherboards

non-data area when not in use to prevent possible

contain two PATA controllers, each capable of

damage to the disk surface. Today's drives use a

supporting up to two PATA devices. By looking at

linear or voice coil motor consisting of a permanent

the motherboard itself or at the motherboard book,

magnet surrounding a coil on the actuator arm.

you can determine which is the primary controller

Electrical current causes the coil to generate a

and which is the secondary. If you are using only

magnetic field that moves the actuator arm and

one controller, it should be the primary one.

thus the read/write heads. Containing no data, one

The Advanced Technology Attachment Packet

■

side of one platter is used as a map to position the

Interface (ATAPI) enables nonhard drive devices

heads directly over the data. Voice coil technology

to use a PATA controller. ATAPI devices, such as

automatically parks the heads when the drive loses

CD-ROM drives, use the same 40-pin interface and

power.

follow the same rules of master, slave, and cable

select jumper settings. Nonhard drives must get

Disk geometry for a particular hard drive consists of

■

their BIOS from option RAM or a software driver.

three primary values: heads, cylinders, and sectors

per track. There are two read/write heads per

Originally, IDE drives used the same BIOS command

■

platter. A hard drive can have either an even or an

set introduced years earlier. Maximum values of

odd number of heads. A cylinder defines a group of

1024 cylinders, 16 heads, and 63 sectors per track

tracks of the same diameter. Each track is sliced into

limited an IDE drive's capacity to 528 million bytes

tiny slivers called sectors, each of which stores

(504 MB). Western Digital developed the LBA

512 bytes of data. Disk geometry uses the number of

sector translation method to accommodate larger

sectors per track. Combining cylinders, heads, and

EIDE drives. LBA supports drives with up to

sectors per track is referred to as CHS. Write

256 heads, for a storage capacity limit of 8.4 GB.

precompensation and landing zone, two other

138

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

As drive capacity neared the 8.4-GB maximum,

All SCSI devices can be divided into two groups:

■

Phoenix Technologies broke the limit by coming up

internal and external. Internal SCSI devices connect

with a new set of BIOS commands called Interrupt

to the host adapter with a 68-pin ribbon cable.

13 extensions (INT13). Completely ignoring the CHS

External devices connect to the host adapter with

values, INT13 supports drives up to 137 GB by

either a 50-pin HD connector or a 68-pin HD

reporting a stream of "addressable sectors" to LBA.

connector. SCSI enables you to daisy-chain devices

together to form longer SCSI chains.

As drive capacity neared the 137-GB limit, the

■

ANSI ATA committee adopted ATA/ATAPI-6, a

Improper termination or incorrect SCSI ID settings

■

new standard that increased the limit to more than

are the two most common causes of SCSI devices

144 petabytes (144,000,000 GB). It uses a 48-bit

not working. The key factor here is that you must

addressing scheme and an enhanced block mode

terminate only the ends of the SCSI chain.

that transfers up to 65,536 sectors at a time.

Protecting Data with RAID

Newer hard drives use direct memory access

■

Drive mirroring writes data simultaneously to two

(DMA) mode to send data directly to RAM,

■

hard drives, enabling the system to continue to

bypassing the CPU. Instead of using the slow

work if one hard drive dies. A faster and even

DMA controller chip, today's DMA transfers use

more effective technique is drive duplexing, which

bus mastering to transfer 32 bits of data. Hard

performs mirroring using separate controllers for

drives typically use one of the following modes:

each drive. A third way to create redundant data is

Ultra DMA mode 4 (ATA/66), Ultra DMA mode 5

disk striping with parity. This technique, requiring

(ATA/100), and Ultra DMA mode 6 (ATA/133).

at least three drives, combines the redundancy of

To use Ultra DMA, you must have a controller and

disk mirroring with the speed of disk striping.

an 80-wire ribbon cable. Some motherboards

Although disk striping without parity works very

combine different speeds of ATA controllers, with

fast, splitting the data across two drives means

the higher speed controller indicated with a bright

you'll lose all data if either drive fails.

color. The 80-wire ribbon cable indicates where the

master and slave drives should be connected.

Numbered 0 through 6, there are seven official

■

High-end PATA devices can use lower end

levels of RAID, but the most commonly used ones

controllers, but they will operate at the slower

are RAID 0 (disk striping), RAID 1 (disk mirroring

speed.

or duplexing), and RAID 5 (disk striping with

distributed parity).

Serial ATA (SATA) devices look identical to

■

standard PATA devices (except their data and

RAID may be implemented through hardware or

■

power connectors), but their thinner seven-wire

software methods. While software implementation is

cables provide better airflow and may be up to a

cheaper, hardware techniques provide better

meter (39.4 inches) long. SATA does away with the

performance. Windows 2000 Server and Windows

entire master/slave concept. Each drive connects to

2003 Server include built-in RAID software for

one port, so no more daisy-chaining drives.

RAID 0, RAID 1, and RAID 5 for either ATA or SCSI.

Windows 2000 and XP Professional include Disk

SATA devices are hot-swappable, great for RAID

■

Management for RAID 0. RAID software solutions

technology. SATA transfers data in serial bursts,

tend to overwork your operating system, resulting in

for up to 30 times the throughput of PATA. SATA

slowdowns. Hardware RAID is invisible to the OS and

drives come in two common varieties, the 1.5Gb

is usually hot-swappable. A hardware ATA RAID

and the 3Gb, that have a maximum throughput of

controller usually requires CMOS configuration. Many

150 MBps and 300 MBps, respectively.

motherboards include built-in ATA-based hardware

RAID 0 and RAID 1 capabilities.

SCSI: Still Around

All SCSI chains, which support either 8 or 16

SCSI drives were a natural for the multiple-disk

■

devices including the controller, require proper

RAID. Specialized ATA RAID controller cards

termination to prevent signal echo. Additionally,

support ATA RAID arrays of up to 15 drives. With

each device on a chain requires a unique SCSI ID,

its hot-swap capabilities, SATA may soon take over

which can be set by jumpers, switches, or dials.

lower end RAID from SCSI.

139

Chapter 8: Hard Drive Technologies

Connecting Drives

autodetection does not provide true BIOS support.

You must still install drivers to provide the BIOS. If

Older PATA drives use a 40-wire cable, while the

■

the driver is installed in a graphical mode, you will

newer Ultra DMA drives use an 80-wire cable.

be unable to access the drive if you boot to a

Either round or flat and containing no twists, each

command promptonly environment.

ribbon cable supports two drives. A diagram on

the hard drive's housing shows how to set its

It takes four things to get a drive installed and

■

jumpers to identify it as master, slave, standalone

recognized by the system: jumpers, data cable,

(on some drives), or cable select (cable position

power, and CMOS setup or providing BIOS. If any

determines whether the drive will be master or

of these steps is missed or messed up, your drive

slave). Two devices on one cable must both be set

simply doesn't exist according to the PC.

to cable select, and the cable itself must also be

If the autodetection feature of the CMOS utility

■

cable select, as indicated with a pinhole through

does not detect a drive, it means it is installed

one wire. Align the colored stripe on the cable with

incorrectly or the drive itself is bad. Check the

pin 1 on the controller and the drive. Use a Molex

master/slave jumper settings. Make sure that the

connector to provide power to the drive.

ribbon cable aligns pin 1 with pin 1, and that the

SATA supports only a single device per controller

Molex connector is supplying power to the drive.

■

channel, so there are no master, slave, or cable

Once you've checked the physical connections, run

■

select jumpers. SATA controller and power cables

through these issues in CMOS. Is the controller

are keyed to prevent incorrect insertion. You can

enabled? Is the storage technology--LBA, Large,

connect a PATA device to a SATA motherboard

INT13, ATA/ATAPI-6--properly set up? What

controller by way of a SATA bridge.

about the data transfer settings for PIO and DMA

modes? Similarly, can the motherboard support the

BIOS Support: Configuring CMOS

type of drive you're installing? If not, you can flash

and Installing Drivers

the BIOS or get a hard drive controller that goes

While system BIOS supports built-in PATA

■

into an expansion slot. Make certain with

controllers, hard drives require configuration in

non-integrated hard drive controllers, such as

CMOS. ATAPI devices require software drivers to

those that come with many SATA drives, that

provide BIOS support. Built-in SATA controllers

you've installed the proper drivers for the

on a motherboard also require software drivers, as

controller. Always check the manufacturer's Web

does a SATA controller on a separate expansion

site for new drivers.

card. All SATA devices get BIOS support from the

SATA controller, but some drives require

Troubleshooting Hard Drive Installation

additional configuration; in particular, with RAID

It takes four things to get a drive installed and

■

systems, you may also have to configure the

recognized by the system: jumpers (PATA only),

controller Flash ROM settings for the specific

data cable, power, and CMOS setup recognizing

drive(s) you install.

the drive. If any of these steps is missed or messed

When the hard drive type is set to "Auto," PATA

up, you have a drive that simply doesn't exist

■

devices can be queried directly by BIOS routines,

according to the PC! To troubleshoot hard drives,

resulting in the correct CMOS settings for up to

simply work your way through each step to figure

four ATA devices. Autodetection made hard drive

out what went wrong.

types obsolete. Because PATA drives have CHS

Once you've checked the physical connections, run

■

values stored inside them, the BIOS routine, when

through these issues in CMOS. Is the controller

set to "Auto," updates the CMOS each time the

enabled? Is the storage technology--LBA, INT13,

computer boots. An alternative is to run the

ATA/ATAPI-6--properly set up? Similarly, can

autodetection option from the CMOS screen.

the motherboard support the type of drive you're

Even if the autodetect feature indicates that an

installing? If not, you have a couple of options. You

■

optical ATAPI drive has been installed, this merely

can flash the BIOS with an upgraded BIOS from

shows that the drive is connected properly and has

the manufacturer or you can get a hard drive

the option to function as a boot device. This

controller that goes into an expansion slot.

140

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

■ Key Terms

40-pin ribbon cable (109)

geometry (105)

SCSI chain (120)

80-wire cable (116)

heads (105)

SCSI ID (122)

Advanced Technology Attachment

integrated drive electronics

sector (106)

Packet Interface (ATAPI) (114)

(IDE) (108)

sector translation (113)

ATA/133 (118)

Interrupt 13 (INT13)

sectors per track (107)

extensions (116)

ATA/ATAPI-6 or simply

Self-Monitoring, Analysis, and

ATA-6 (117)

logical block addressing

Reporting Technology

(LBA) (112)

cylinder (106)

(S.M.A.R.T.) (115)

PIO modes (111)

disk duplexing (124)

serial ATA (SATA) (108)

parallel ATA (PATA) (108)

disk mirroring (124)

small computer system interface

(SCSI) (120)

Partial Response Maximum

disk striping (124)

Likelihood (PRML) (103)

stepper motor (104)

disk striping with parity (126)

redundant array of independent

termination (123)

DMA modes (111)

(or inexpensive) disks

track (106)

Enhanced IDE (EIDE) (111)

(RAID) (126)

Ultra DMA (116)

External SATA (eSATA) (120)

SATA bridge (119)

voice coil (104)

■ Key Term Quiz

Use the Key Terms list to complete the sentences that

5. Seen in RAID 5, ____________ uses at least three

follow. Not all terms will be used.

drives and combines the best features of disk

mirroring and disk striping.

1. An ATA hard drive connects to the controller

6. A(n) ____________-compliant CD-ROM drive

with a(n) ____________ while an Ultra DMA

installs and cables just like an EIDE drive.

drive uses a(n) ____________.

7. The ANSI ATA committee adopted the

2. A(n) ____________ is composed of a group of

____________ standard, called "Big Drives" by

tracks of the same diameter that the read/write

Maxtor, that allows drives with more than 144

heads can access without moving.

petabytes.

3. To install a parallel ATA device to a serial ATA

8. ____________ drives transfer data at 133 MBps.

controller, use a tiny card called a(n)

9. Drives that use ____________ bypass the CPU

____________.

and send data directly to memory.

4. LBA, developed by Western Digital, uses

10. ____________ devices require termination at both

____________ to get around the limits of 1024

ends of a chain.

cylinders, 16 heads, and 63 sectors/track.

■ Multiple-Choice Quiz

1. Which of the following is NOT used to compute

2. Which level of RAID is disk striping with

storage capacity in CHS disk geometry?

distributed parity?

A. Sectors per track

A. RAID 0

B. Tracks

B. RAID 1

C. Heads

C. RAID 5

D. Cylinders

D. RAID 6

141

Chapter 8: Hard Drive Technologies

A. IDE

3. Which of the following is the most efficient

encoding method?

B. EIDE

A. Partial Response Maximum Likelihood (PRML)

C. SCSI

B. Frequency modulation (FM)

D. ATA

C. Run length limited (RLL)

9. Shelby wants to add a new 100-GB hard drive to

her computer. Which of the following will allow

D. Modified frequency modulation (MFM)

her to do so?

4. Counting both channels, what is the maximum

A. CHS

number of drives/devices that EIDE can

support?

B. LBA

A. One

C. ECHS

B. Two

D. INT13

C. Seven

10. Which of the following techniques provides

redundancy by using two disks and two

D. Four

controllers?

5. Which of the following is NOT true about cable

A. Drive mirroring

select?

B. Drive duplexing

A. Both drives/devices should be set for cable

select.

C. Disk striping

B. It requires a special cable with a pinhole

D. Disk striping with parity

through one wire.

11. How many wires does an Ultra DMA cable

C. The colored stripe on the ribbon cable should

have?

align with pin 1 on the controller and drive.

A. 24

D. Position of the drives on the cable does not

B. 34

matter.

C. 40

6. If you install two IDE drives on the same cable,

D. 80

how will the computer differentiate them?

12. Billy just installed a second hard drive, but the

A. The CMOS setup allows you to configure

autodetection utility in CMOS does not detect it.

them.

Sara told him he probably had the jumpers set

B. You must set jumpers to determine which

incorrectly or had forgotten to connect the Molex

drive functions as master and which

power connector. John told him his new hard

functions as slave.

drive is probably bad and he should return it. Is

C. You will set jumpers so each drive will have a

Sara or John probably correct?

unique ID number.

A. Sara is correct.

D. The drives will be differentiated by whether

B. John is correct.

you place them before or after the twist in the

C. Neither is correct.

ribbon cable.

D. Either John or Sara may be correct.

7. What was the maximum hard drive size allowed

13. Which of the following is NOT an advantage of

by BIOS routines for the original AT command

serial ATA (SATA)?

set?

A. It is hot-swappable.

A. 528 MB

B. Thinner cables provide better airflow inside

B. 1024 MB

the case.

C. 504 MB

C. SATA provides faster data throughput than

D. 1028 MB

PATA.

8. Which of the following terms does NOT describe

D. SATA cable must be shorter than PATA

parallel ATA devices?

cables.

142

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

14. Which of the following two CMOS configuration

15. What standard did the ANSI ATA committee

options are obsolete with today's hard drives?

adopt that increased disk storage capacity to

more than 144 petabytes?

A. Cylinders and heads

A. ATA/ATAPI-6

B. Heads and sectors

B. LBA

C. Sectors and write precompensation

C. INT13

D. Write precompensation and landing zone

D. ECHS

■ Essay Quiz

1. Discuss at least three advantages of serial ATA

features including storage capacity, interface,

over parallel ATA.

RPM, and cost:

2. Compare and contrast hardware and software

Maxtor Model # 6Y120L0 and Maxtor Model

■

RAID implementation.

# 6Y120M0

3. Your friend Blaine has a Pentium III computer

Western Digital Model # WD1200JB and

■

with a 100-MHz bus. Currently, it has only a

Western Digital Model # WD1200JD

20-GB ATA/100 hard drive and a CD-RW drive.

Since he's interested in graphics, he knows he

5. Hard drives include other features and

needs more storage capacity and wants to add a

characteristics not included in this chapter.

second hard drive. What advice will you give

Choose one of the following topics and use the

him about selecting a new hard drive?

Internet to define and explain it to the class.

4. Use www.google.com or a site such as

Zone bit recording

■

www.newegg.com to compare one of the

"Pixie dust" hard drives

■

following pairs of hard drives to determine their

Lab Projects

· Lab Project 8.1

Access the CMOS setup for your computer and

utility and run it. Does it offer different modes with

examine the settings that apply to your hard drive(s)

different drive capacities? Try to find a screen that

and EIDE interface. In particular, look at the initial

includes PIO modes and examine this setting. What

screen to see if it is set to autodetect the kind of hard

other screens apply to the hard drive? When you

drive. Is the mode set to LBA or something else?

finish, be sure to choose Quit without Saving.

Now find the screen that includes the autodetect

· Lab Project 8.2

Visit your local computer store or use the Internet to

parallel ATA interfaces or if they offer serial ATA

discover what kinds of hard drives and hard drive

interfaces, either onboard or through an expansion

interfaces are commonly offered with a new computer.

card. If you were purchasing a new computer, would

Try to determine whether the motherboards offer only

you select PATA or SATA? Why?

143

Chapter 8: Hard Drive Technologies

· Lab Project 8.3

Your supervisor has decided to implement RAID on

security of data at the lowest cost possible, and the

the old company server machine, where everyone

other that maximizes speed but retains some

stores their work-related data. Come up with two

security. Cost is not a factor for the second RAID

competing RAID setups, one that maximizes

plan.

· Lab Project 8.4

If your lab has the equipment, install a second hard

detected. Boot into your operating system and verify

drive in your system. Install it on the same cable as

that both drives are accessible. (You may need to

the existing drive and jumper it as the slave. (You

partition and format the second drive before it is

may need to jumper the existing drive as master.)

actually usable.)

Reboot, enter CMOS, and verify both drives are

144

Mike Meyers' CompTIA A+ Guide: PC Technician (Exams 220-602, 220-603, & 220-604)

Table of Contents: