|
Hard Disk Basics
Definition of HDD / hard disk / hard drive
Hard Disk Drive is one of several types of magnetic media used for storing data. Unlike floppy disks, hard
disks are non-flexible and non-removable. They hold much more data than floppy disks, and are the principle long-term
data storage for most personal computers. Non-removable hard disks are also known as "fixed disks".
Hard Disk Drives are mechanical with very sophisticated parts. Typically, a safe hard disk life-span is
considered to be 5-years for home computing. Server- and Workstation hard disks have to be re-placed more often
as they are operating on a 24/7 basis.
Hard Disk is basically not very different from a cassette tape. Both hard disks and cassette tapes use the
same magnetic recording techniques. Hard disks and cassette tapes also share the major benefits of magnetic storage,
the magnetic medium can be easily erased and rewritten, and the stored media with all the magnetic flux patterns
stays stored onto the medium for many years.
Hard Disk Performance is measured in the speed that the platters spin and data is written / read. Hard disk
spindle speeds can be from 5400 RPM, 7200 RPM, 10000 RPM or 15000 RPM. Notebook hard disks perform at 4200 RPM,
5100 RPM and 7200 RPM spindle speeds.
| Hard Disk |
|
Tape |
| Hard Disk is a self-contained unit with built-in Read-Write
Heads and magnetic platters. |
|
There are two parts to any tape magnetic recording system:
the recorder itself (which also acts as the playback device) and the tape it uses as the storage medium. |
| The magnetic recording material is layered onto a high-precision
aluminum or glass disk. The hard-disk platter is then polished to mirror-type smoothness. |
|
The magnetic recording material on a cassette tape is coated
onto a thin plastic strip. |
| In a hard disk, the read/write head "flies" over
the disk, never actually touching it. |
|
In a cassette-tape deck, the read/write head touches the
tape directly. |
| Move to any point on the surface of the disk almost instantly. |
|
Fast-forward or reverse to get to any particular point on
the tape. This can take several minutes with a long tape. |
| A hard-disk platter can spin underneath its head at speeds
up to 3,000 inches per second (about 170 mph or 272 kph). |
|
The tape in a cassette-tape deck moves over the head at
about 2 inches (about 5.08 cm) per second. |
The information on a hard disk is stored in extremely small magnetic domains compared to a cassette tape's. The
size of these domains is made possible by the precision of the platter and the speed of the medium.
Because of these differences, a modern hard disk is able to store an amazing amount of information in a small space.
A hard disk can also access any of its information in a fraction of a second.
Basic Handling Precautions:
Never drop, jar, or bump the drive.
DO NOT connect/disconnect any drive cables when the power is on.
DO NOT force or rock the connectors into their sockets. Push them in straight until they are seated.
Allow the drive to reach room temperature before installing it in the computer.
Helpful hint - to avoid electrostatic discharge(ESD) damage touch the metal case of your PC system or use
a grounding strap before handling. Connect your grounding strap to the system casing.
back to
contents
All hard disks consist of thin platters with
a magnetic coating. They rotate quite fast inside a metal container. Data are written and read by read/write heads,
which are designed to ride on a microscopic cushion of air, without touching the platter. They register bits from
the magnetic coating, which races past them.
| Magnetic Recording Basics |
|
|
 |
|
The drive channel electronics receive data in binary form
from the computer and convert them into a current in the head coil. The current in the coil reverses at each 1
and remains the same at each 0.
This current interaction with the media results in magnetization of the media, which direction depends on the current
direction in the coil. |
| This figure illustrates a writing sequence. |
|
|
The reading process includes excitation of the current in the head coil when the head "senses"
changes in the magnetic flux. The read voltage pulses at the flux transitions are then translated into sequences
of bits equal to 0 and 1. The so-called Wallace's spacing loss factor postulates that the loss of magnetic signal
power will be proportional to the media - head separation. This requires magnetic heads to fly as close to the
disk surface as possible, which forces modern heads to fly at a few nanometers only.
| Magnetic Recording Heads |
|
|
 |
|
Magnetic head typically consists of an MR (magneto-resistive)
or GMR (giant MR) reading head and a thin-film inductive write head.
MR head design is based on the ability of metals to change their resistivity in the presence of a magnetic field.
The alloy of Ni and Fe (81%/19%) is widely used in MR heads and is called Permalloy.
MR heads are suitable for extremely high bit density and have superior signal-to-noise ratio when compared
to inductive read heads.
Inductive thin film heads generate strong magnetic fields at the gap between the poles, thereby magnetizing
areas of the media. |
This figure illustrates MR Reading and
Inductive Writing Heads. |
|
|
back to
contents
On the illustration below, you see a hard disk with three platters. It has 6 read/write heads, which move synchronously.
| Drive Physical and Logical Organization |
|
|
 |
|
Track
A concentric set of magnetic bits on the disk is called a track. Each track is divided into 512 bytes (usually)
sectors.
Sector
A part of each track defined with magnetic marking and an ID number. Sectors have a sector header and an ECC (Error
Correction Code). In modern drives, sectors are numbered sequentially.
Cylinder
A group of tracks with the same radius is called a cylinder (red tracks on the picture belong to one cylinder).
Data addressing
There are two methods for data addressing: CHS (Cylinder-Head-Sector) and LBA (Logical Block Address). CHS is used
on most IDE drives, while LBA is used on SCSI and enhanced IDE drives. |
| This figure illustrates Physical and Logical Organization
of a drive. |
|
|
Arms, which guide the movement of the read/write heads, move in and out. As illustrated, there will typically
be 6 arms, each with read/write heads. The synchronous movement of these arms is performed by an electro-mechanical
system called the head actuator. The hard disk data can only be attained via one head at a time.
back to
contents
Read / Write Head consists of a tiny electromagnet. The shape of the head end acts like an air foil, lifting
the read/write head slightly above the spinning disk below.
When the disk rotates under the read/write head, it can either read existing data or write new ones:
Write mode: Current is applied to the coil, the head will become magnetic. This magnetism will orient the
micro magnets in the track.
Read mode: When the head moves along the track without current applied to the coil, it will sense the micro
magnets in the track. This magnetism will induce a current in the coil. These flashes of current represent the
data on the disk.
The read/write heads are by far the most expensive parts of the hard disk. They are incredibly tiny. In
modern hard disks they float between 5 and 12 micro inches (millionths of an inch) above the disk. When the PC
is shut down, they are auto parked on a designated area of the disk, so they will not be damaged during transport.
back to
contents
CHS addresses data by simply specifying the cylinder (radius), head (platter side), and sector (angular
position).
LBA assigns each sector of the drive a sequential number, which is simpler.
If you look into your BIOS, you will find listed the number of cylinders, heads, and sectors for each drive you
have. Modern operating systems access data using LBA directly without the help of the BIOS. This reduces incompatibilities.
To improve performance and increase data rate, HDDs utilize a small amount of fast solid-state memory to store
the most frequently used data. This memory is called 'cache' or 'buffer'. There are two types of cache memory organization:
look-ahead and write / read.
back to
contents
Low Level Formatting an IDE Hard Drive
This is not the actual "low-level-formatting" of the IDE or ATA hard drives. This term is a misnomer
from the old MFM hard drive days when a drive could have the tracks and sectors defined using a Low Level Format.
This is an old term that really does not apply to today's IDE/ATA hard drives.
The IDE or ATA drives used today have all this information preset at the factory and a real Low Level Format would
destroy the drive or at least slow it down radically as it is not possible to redefine the tracks and sectors on
these drives with a true Low Level Format. So, the physical geometry of current IDE/ATA drives cannot be changed
without destroying it. Nowadays Low Level Format is actually redefining the tracks and sectors on a hard drive
and writing the entire disk with zeros.
Most people are using Low Level Format when having a real problem with an IDE hard drive (is referred to as reinitialize
or mid-level format the drive). Usually, one of the following issues has occurred:
- The drive has a boot sector virus.
- The drive has begun to develop numerous bad sectors and they are increasing, (usually seen when running Scan
Disk).
- The drive has had Linux, WindowsNT or another operating system installed creating a Fat System on the drive incompatible
with the new operating system to be installed.
Re-Initializing an IDE Drive
When re-initializing a hard drive, basically a utility is used that over-writes the whole drive with ones and zeros.
Every area of the hard drive is cleansed. These utilities are available from the hard drive manufacturer's websites.
Every drive manufacturer provides such a utility, though some of them have begun to call it a Low Level Format
utility. Once this process is complete, the drive will be void of any partitions. So, it is needed to use FDISK
for Microsoft products to partition the drive and then format the new partition.
back to
contents
Command Queuing
Command queuing enables a hard drive to accept multiple commands from the host controller and rearrange
the completion order of those commands to maximize throughput. The major portion of the drive’s command service
time is seek and rotational delay for the drive head to land on the appropriate data to transfer. The drive can
use rotational optimizations to select the next command to complete so that the major components of the service
time, seek and rotational delay, are minimized.
Rotational optimizations can be achieved when the drive selects the next command based on the fact that
it is closest rotationally to the head's current position. The seek optimizations can be achieved when the drive
selects the next command based on their cylinder number, compared to the cylinder where the heads currently are
located. A major advantage to command queuing is that the command issue and completion overhead may be overlapped
with the drive seek and rotational delay for a different command's data transfer. For example, while a new command
is being issued to the drive, the drive may be seeking to locate the appropriate track on disk for data for a different
command. In essence, the latency for issuing the new command is saved since it was overlapped with the seek for
another command.
Figure below shows the host controller communicating to a hard drive. Each queued command has an associated tag
value. The host in both the data transfer phase and completion phase of a queued command uses the tag. To transfer
data for a particular command, the disk drive communicates to the host the tag of that command. The host then sets
up a DMA operation that points to the appropriate host memory region for that command. To complete a command, the
disk drive sends the associated tag of the command to be completed. The use of tags allows the host to protect
memory regions from being errantly accessed by the device. The drive also sends status information to update the
host controller with the status of pending command.
| Command Queuing Block Diagram |
|
|
back to contents
Drive Performance Issues
There are ways to measure the performance of a hard disk. Seek time along is not enough to indicate the
preformance of a drive. Seek time is affected by the technology used to position the heads on the disk: Either
a Band Stepper or Voice Coil.
SEEK TIME: Average time to locate a cylinder on the drive typically 8 to 16 mS.
ACCESS TIME: Average time to locate a cylinder and a specific sector on the drive and begin typically
10 to 30 mS.
ACCESS TIME = SEEK TIME (Time to move to the cylinder) + ROTATIONAL LATENCY TIME (Time to wait for sector)
The seek time is the amount of time between when the CPU requests a file and when the first byte of the file is
sent to the CPU. Times between 10 and 20 milliseconds are common. This is affected by Rotational Speed: at 3600
rpm, Rotational Latency is 8.33mS. Newer drives, rotating at 7200 rpm have faster access times.
DATA TRANSFER RATE: The data rate is the number of bytes per second that the drive can deliver to the CPU.
Rates between 5 and 40 megabytes per second are common.
INTERLEAVE is only used for OLDER drives, not IDE or SCSI and hence is irrelevant for most modern applications.
Capacity - The other important parameter is the capacity of the drive, which is the number of bytes it can
hold.
back to
contents
Basically, the more clusters on your hard drive, the slower the performance. This is true for any file system,
FAT16, FAT32, NTFS, HPFS, whatever. However, since FAT32 allows for many more clusters on a single partition than
FAT16, the effect may be noticable. Disk utilities are especially affected (slower) the more clusters in the partition.
For instance, the closer to 8GB your partition gets, the more 4K clusters, and the slower the performance. (However,
as your partition gets bigger, your slack vs. FAT16 improves.)
In preliminary benchmark testing, FAT32 and FAT16 benchmark roughly the same (within 2% either way) when partition
size and cluster size are the same. (Note that in order to create a FAT32 partition with the same cluster size
as FAT16, the /Z switch must be used when formatting.) However, when cluster sizes are made smaller and the number
of clusters increases (only possible with FAT32), disk performance degrades.
Thus there is a battle between slack and performance: Small clusters mean less slack but worse performance.
Large clusters mean more slack but better performance. Most users will notice slack differences much more
than performance differences caused by varying cluster sizes. Microsoft has decided for us (in making 4K the default
cluster size for FAT32) that 4K clusters is the best balance between slack and performance. However, with the /Z
switch on the FORMAT command, the user has the ability to decide for his/herself what cluster sizes should be,
based on the user's concerns about slack vs. performance.
Read the table below for choosing the cluster size on basis of disk size for better performance and limiting slack
space.
Best recommended Cluster size for your drive
| Disk Size |
Cluster Size Recommended |
Format Command Option* |
| Less than 1 GB |
4 KB
|
Format x: |
| Less than 4 GB |
8 KB
|
Format x: /z:16 |
| Between 4 to 32 GB |
32 KB **
|
Format x: /z:64 |
*x: is the drive i.e. e.g. for c drive it is format
c:
**32 KB Cluster Size is best for performance, even faster than FAT16. Test drive runs on FAT32 with cluster
size of 4 KB(4096bytes) and System Information by Norton Utilities 4.0 gives the cached read/write benchmark of
102.7 MB and 53.3 MB physical read/write speed respectively on 30GB Maxtor 33073H3.
Abbreviations used in the following data-tables:
| ATA |
AT attachment |
|
MB |
megabyte |
| bpi |
bits per inch |
|
Mbits/sec |
megabits per second |
| CHS |
cylinder - head - sector |
|
MB/sec |
megabytes per second |
| db |
decibels |
|
MHz |
megahertz |
| dbA |
decibels, A weighted |
|
ms |
millisecond |
| DMA |
direct memory access |
|
MSB |
most significant bit |
| ECC |
error correction code |
|
mV |
millivolts |
| fci |
flux changes per inch |
|
ns |
nanoseconds |
| G |
acceleration |
|
PIO |
programmed input/output |
| GB |
gigabyte |
|
RPM |
revolutions per minute |
| Hz |
hertz |
|
tpi |
tracks per inch |
| KB |
kilobyte |
|
UDMA |
ultra direct memory access |
| LBA |
logical block address(ing) |
|
µsec |
microsecond |
| LSB |
least significant bit |
|
V |
volts |
| mA |
milliamperes |
|
W |
watts |
Sources: Symantec Corp., Microsoft, Maxtor, IBM, KEPCIL
Designs.
| Hard disk drive: |
30GB Maxtor 33073H3 |
| Firmware: |
YAH814Y0 |
| Hard disk drive size: |
3.5 in |
back to
contents
Performance Specifications
| MAXTOR MODELS |
34098H4 | 33073H3 | 32049H2 | 31024H1
|
| Seek Times (typical read) |
| Track-to-Track |
1.0 ms |
| Average (performance) |
9.5 ms |
| Average (silent mode) |
15 ms |
| Full Stroke |
< 20.0 ms |
| Average Latency |
5.55 ms |
| Rotational Speed (+ 0.1%) |
5400 RPM |
| Controller Overhead |
< 0.3 ms |
Data Transfer Rate
To / From Interface
(Ultra ATA / 100, DMA - M5) |
up to 100 MBytes / sec. |
To / From Interface
(P IO 4 / Multi-word DMA M5) |
up to 16.7 MBytes / sec. |
| To / From Media |
up to 46.7 MBytes / sec. |
| Start Time (0 to Drive Ready) |
8.5 sec. typical |
Drive Configuration
| MODEL |
Maxtor 33073H3 |
| Integrated Interface |
ATA-5 / Ultra ATA/100 |
| Encoding Method |
E2 PR4 RLL 16/17 |
| Interleave |
1:1 |
| Servo System |
Embedded |
| Buffer Size / Type |
2 MB SDRAM |
| Data Zones per Sur face |
16 |
| Data Surfaces / Heads |
3 |
| Number of Disks |
2 |
| Areal Density |
14.7 Gbits / in2 max |
| Track Density |
34 000 tpi |
| Recording Density |
354 to 431 kbpi |
| Bytes per Sector / Block |
512 |
| Sectors per Track |
373 to 746 |
| Sectors per Drive |
60 032 448 |
Performance Specifications
| MODEL |
Maxtor 33073H3 |
| Seek Times (typical read) |
|
| Track- to-Track |
1.0 ms |
| Average (performance) |
9.5 ms |
| Average (silent mode) |
15 ms |
| Full Stroke |
<20.0 ms |
| Average Latency |
5.55 ms |
| Rotational Speed (±0.1%) |
5400 RPM |
| Controller Overhead |
< 0.3 ms |
| Data Transfer Rate |
|
| To/From Interface (Ultra ATA/100, DMA - M5) |
up to 100 MBytes/sec |
| To/FromInter face (P IO 4/Multi -word DMA M5) |
up to 16.7 MBytes/sec |
| To/From Media |
up to 46.7 MBytes/sec |
| Start Time (0 to Drive Ready) |
8.5 sec typical |
back to contents
About diagnosing and fixing problems
I am using Symantec Norton Utilities to "keep my computer running". It has disk utilities to test, diagnose
and repair hard disk errors. Trouble-free computing depends on the integrity of your computer, which requires an
error-free hard disk and a correctly installed copy of Windows. The trouble is, although most computers start out
this way, over time Windows and hard drives are likely to degrade. This degradation, if not corrected, can ultimately
lead to data loss. The best cure for any problem is prevention. Many Norton Utilities programs are on alert for
the most common error conditions that can cause computer lock-ups and crashes, and in turn lead to computer problems
and data loss. If a Norton Utilities program detects a potential problem, it can correct the condition automatically,
or warn you, giving you the option to fix the problem if possible.
Well, actually I have downloaded some other additional utilities, such as EasyCleaner from ToniArts, SpyBot's Search
& Destroy and of course always doing the old way, manually checking and cleaning the registry with Windows
Registry Editor which I prefer the most. Anyway, nowadays there are many "background" programs that are
affecting the overall performance of the hard disk. You should manually configure your programs to give permit
to access the hard disk resources while another program is running.
Sample Test Result:
Disk Doctor
Norton Utilities for Win95
June 26, 2001 8:52am
*************************
* Report for Drive C: *
*************************
DISK TOTALS
----------------------------------------
6,000,128 kilobytes Total Disk Space
1,690,210,304 bytes in 13,426 User Files
4,349,952 bytes in 987 Directories
15,572,992 bytes in 691 Hidden Files
4,318,208 kilobytes available on the disk
LOGICAL DISK INFORMATION
----------------------------------------
Media Descriptor: F8
Large Partition: Yes
FAT Type: 32-bit
Total Sectors: 12,000,492
Total Clusters: 1,497,132
Bytes Per Sector: 512
Sectors Per Cluster: 8
Bytes Per Cluster: 4,096
Number of FATs: 2
First Sector of FAT: 32
Number of Sectors Per FAT: 11,700
First Cluster of Root Dir: 2
Number of Clusters in Root Dir: 1
First Sector of Data Area: 23,432 |
PHYSICAL DISK INFORMATION
----------------------------------------
Drive Number: 80
Heads: 255
Cylinders: 3,736
Sectors Per Track: 63
Starting Head: 1
Starting Cylinder: 0
Starting Sector: 1
Ending Head: 254
Ending Cylinder: 746
Ending Sector: 63
SYSTEM AREA STATUS
----------------------------------------
No errors in the system area
FILE STRUCTURE STATUS
----------------------------------------
No errors in the file structure
FREE SPACE STATUS
----------------------------------------
No errors in the free cluster count
SURFACE TEST STATUS
----------------------------------------
Test Settings
-----------------------
Test: Entire Disk Area
Test Type: Normal Test
Repair Setting: Prompt before Repairing
Passes Requested: 1
Passes Completed: 1
Elapsed Time: 16 minutes, 7 seconds
No errors encountered in Surface Test |
back to contents
DiamondMax Plus 9
The 7200 RPM DiamondMax Plus 9 hard drive is your solution for performance PCs with data- intensive applications
such as network attached storage, home networking devices, audio, video and other multimedia applications.
- Formatted capacity: 60, 80,120, 160 and 200GB
- Average seek time: =9.3
- Rotational speed: 7200 RPM
- 2MB or 8MB cache buffer
- Fast ATA/Enhanced IDE Compatible
- Ultra ATA/133 Data Transfer Speed
- Serial ATA version enables transfer speeds up to 150 MB/sec
- 2MB and 8MB Cache Buffer
- Quiet Drive Technology
- 100 % FDB (fluid dynamic bearing) motors
- Maxtor Shock Protection System
- Maxtor Data Protection System |
More Details from Maxtor...
|
back to TOP
Any questions or inquiries may be e-mailed to keppanet@hotmail.com.
|
