| |
This page faces some technical matters that concern the hard disk and
the rescue of the data on you support magnetic in general.
I list matters essays:
As you/he/she is done inside a hard disk?
Here is an open disk!
Here
above a disk IDE of the last generation is represented, deprived
of the protective cover (for the most curious it deals with a Maxtor
DiamondMax from 3.5", 5400 rpms and SMART). Also the disks SCSIs
are constructively similar and you/they can be brought principal
parts to this immagine.Le they are contained in a chassis, typically
of light league pressofusa, closed by a cover to watertight estate,
endowed with gaskets and of a lot of grapevines of fixing. To the
inside:
-
the
disks: in the model in figure they are four, but you/they can
be from 2 in on
-
the
principal motor (not visible, because on the opposite side),
that brings in rotation the disks
-
the
heads, set to the extreme of an arm it brings head: I am a couple
for every disk and they act around two faces
-
the
motor of movimentazione, that makes to complete a movement along
an arc of circle to the arms handed heads
-
the
electronic circuit of inside management of the signals from
and for the heads, currently constituted by an alone devoted
chip
-
the
inside wiring: in this case constituted by a flexible printed
circuit
-
the
connector IDE to 40 poles, standard (or standard SCSI to 50
or wide to 68 poles)
-
the
connector of feeding to 4 poles, standard
-
the
jumperses of configuration master/slave and for the various
options foreseen by the Builder
-
the
printed circuit of the controller (not visible, because on the
opposite side)
As
you/he/she can be seen, the constructive parts of the disk are ridottissime
and the operation it is insured from the elevated technology of
the materials employees and from the experience that you/he/she
is developed in these years and that you/he/she has brought to the
actual results. Shortly time has passed by the by now prehistoric
10 great MB as a box from shoes, with times of access of a lot of
about ten milliseconds to the 36 GBs in the format 3.5" with inferior
times of access to the 9ms. To es., Maxtor, in one publicity of
his, it notices as 5 years his/her productive lines they were able
of sfornare a disk from 2 GBs every 9 seconds and today is stopped
to a disk by 20 GBs every 2 seconds. For this the Builders assure
periods of guarantee of 2 or more years, up to 5 or 6 or more for
the SCSIs. However it is had to notice that the mortality of the
disks, especially in the initial phase of job (childish mortality)
it is still well present, even if a lot of redoubt in relationship
to the past and the concept of backup must be held well present,
if it is desired to work in safety.
Also particular at first sight not evident they are dependent from
a very sophisticated technology. To es.., inside the chassis an
is introduced' checked atmosphere and the whole assemblage of the
parts is effected in deprived rooms of dust, said white rooms, where
the percentage of dust has to be inferior to few parts for million
(against the many million of the common environment! ). This checked
atmosphere is an essential factor for the operation of the disk
as the microscopic heads sorvolano to elevated speeds the delicate
surfaces of the dishes, without touching her, to few milionesimis
of millimeter, effects graces aerodynamic and the presence of the
smallest grain of impurity would be harmful as for an auto in run
the to be himself/herself/themselves before a boulder. Insofar
the opening of the protective cover, in normal environment, even
to give a' glance to the inside, provokes the immediate death of
the disk (and the to decay some Guarantee!).
GO UP
The motors
The drive contains two motors: one serves to bring in rotation the
dishes, the other to move the heads.
It deals with motors of tall precisone, electronically checked by
special integrated circuits.
The disks ask for a feeding to +5V for the logical circuits and
+12V for the motors. A standard connector (type AMP) you/he/she
has been select in common for all the magnetic device both HDD and
CDROM or similar in the format 51/4." The consumption of the disks
of the last generations is very contained and the most greater part
of the energy is employed really from the motors. The maximum tide
is absorbed during the starting of the disk to bring in rotation
the dishes (spin-up), after that there will be impulses of tide
during the movimentazione of the heads. The least consumption is
had during the state of sleep or stand by, in which the motors and
the part of control are stopped it prepares him under conditions
of least consumption. The following chart brings some values for
a comparison:
spin-up |
operational |
sleep |
10/20W |
5/6W |
0.5/1W |
Then,
in the modern drives, the problem of the electric consumption is
relatively little main point, while it is being him/it that of the
cooling, in how much a certain part of the absorbed energy is dissipated
in heat that must effectively be eliminates, punishment a shortening
of the life of the disk. Despite the employment of materials to
elevated technology, it needs her/it to respect the tolerances micrometriche
and the inside job of the parts, necessary to the attainment of
elevated performances more and more they make the heat a notable
enemy. Up to a little while ago it was common to control's electronics
the to effect a thermal ricalibrazione, necessary to maintain the
correct alignment between the heads and the dishes, following the
due deformations to the heat; this ricalibrazione asks for a momentary
arrest in the transfer of the data, while the controller handles
the operation. It achieves of it that, for many trials in which
the constant flow of data is vital (masterizzazione, audio, digital
video) the ricalibrazione, if extended over a certain time, it constitutes
a notable damage to the integrity of the same trial.
In the most recent disks this king alignment is effected to the
flight (on fly) during the ordinary funzionamentoe and the problem
of the loss of synchronism is not verified among anymore the trials
and the flow of the data.
Currently the standard dimension is 3.5" of width and 1/2" of thickness
and in this format they are available disks IDE up to 26 GBs and
disks SCSI up to 36 GBs, while for these last ones, with greater
mechanical formats, it persuades her you/they can climb to a lot
of about ten GB.
GO UP
We
see in detail some parts:
The printed circuit of the controller
A card contains all the parts that
constitute the controller of the disk. In general you/he/she is
constituted by a printed circuit boxed in the body of the disk and
container a microprocessor, different auxiliary circuits, of the
memory, both for inside use is with functions of cache. The circuit
develops the followings functions:
-
Control of the motor of rotation of the dishes
-
Control of the movement of the attuatore of the heads
-
Management of the interface with the outside (IDE or other)
-
Management of the functions of energetic saving (if you introduce).
-
Management of the functions of correction of the errors, control of the
flow of the data, SMART (if implemented)
Every builder resolves the problem with a different circuit and the same
builder often employs different solutions, while you/he/she is staying
standardized the system of interfacciamento with the rest of the
PC.
The practice they exist different standards for the interface; the principals
are:
It is important to consider that a fundamental difference exists between
interface and system of control of the disk.
The disks with the old interfaces ST506 or ESDI had the situated
controller on the card that was inserted in the bus of the mainboard,
while the situated electronics on the real disk had limited functions
to the writing and the reading of the data following the commands
transmitted by the controller.
In the most recent disks or those said AT-Bus, developed then him
as IDE/EIDE/ATA, instead, the whole logic of control is situated
on the disk and the system it makes only available a channel of
communication constituted by the doors IDE of interface.
The
following image is relative to an old model with interface St-506,
but it sufficiently exemplifies how much I dictate.
The microprocessor of control of the disk is, obviously endowed with a
proper firmware (an inside software of management to the circuit,
at times on on EPROM or on ROM, more rarely in Flash), of an own
Ram, of one or more clocks and of circuits of opportune I/O.
It is often available also a certain quantity of cache that constitutes
a buffer to treat in fluidder way the flow of the data from and
toward the disk, usually from 128, 256 or 512kB. Generally, the
writing and reading of the data on the disks it is very faster than
the transfer of the data themselves through the interface; the cache
constitutes therefore a valid reservoir for the momentary parking
lot of the data. The use of the cache can reduce of very the time
of access to the data. (Attention not to confuse the small cache
on the circuit of control of the disk with the cache that the operating
system can form for the data management of the hard disk and that
you/he/she is drawn in the principal memory of the system). More
before other information on the matter.
 The printed circuit has been
removed by the carcass of the disk and rotated of 90 degrees toward
left.
The containing EPROM the firmware of management of the disk can
be noticed, close to the CPU (a processor serious Hitachi 63xx).
The motors are checked by opportune drives of power that furnish
the necessary signals to the correct operation, under control of
the CPU; a bundle of cables colleague the motors to the printed
circuit. In the photo it is visible connected only the motor of
rotation (spin); the cable of the motor of movimentazione of the
heads was too much short to be able to be connected to its connector
with the rotated circuit.
To notice the sagomatura of the printed circuit, necessary to make
to reenter him/it in the measures typical of encumbrance of the
disk.
 Gives the complexity of the
controller, also the opposite face of the printed circuit is full
of components. The photo brings the circuit riposizionato on the
disk, as it is normally. They notices a discreet quantity of small
component surface mount (to superficial assemblage) and, in the
angle in low to the left, the motor of the heads, that, sticking
out as thickness, you/he/she has forced the planner to shape around
the printed circuit everything; the non rectilinear form of the
edge in some points is due to the necessity to leave space to the
structure of support of the carcass (not inserted in the photo for
greater clarity).
The connector of the interface is of the old type St-506, that requires
two cables toward the card of control it posts on a slot of the
PC; the contacts are electroplated in gold to improve the quality
of the electric connection.
Since control is effected by a microprocessor it is, obviously, necessary
a proper firmware (an inside software of management to the circuit,
at times on on EPROM or on ROM, more rarely in Flash), of an own
Ram, of one or more clocks and of circuits of opportune I/O. It
is often available also a certain quantity of cache that constitutes
a buffer to treat in fluidder way the flow of the data from and
toward the disk, usually from 128, 256 or 512kB. Generally, the
writing and reading of the data on the disks it is very faster than
the transfer of the data themselves through the interface; the cache
constitutes therefore a valid reservoir for the momentary parking
lot of the data. The use of the cache can reduce of very the time
of access to the data. (Attention not to confuse the small cache
on the circuit of control of the disk with the cache that the operating
system can form for the data management of the hard disk and that
you/he/she is drawn in the principal memory of the system). More
before other information on the matter.
GO UP
Interface
The interface is that part of circuit,
hardware and software, that it allows the connection of the disk
the central unity. Types have been implemented different of it,
some of which, with the time, they have become some standards.
Without getting here into greater details, historically the most
important standards have been:
-
ST-506. By now obsolete, it was the interface of the first PC.
The real controller was a card inserted on the bus of the PC
and connected to the disk with two flat cables of different
width. The physical limit is of two unities for controller.
-
XTbus and ATbus. In the first PC XT and AT it was an innovative
structure in which the card on the bus constituted only a door
and the real controller you/he/she was assembled on the disk.
You/he/she immediately is developed in the standard IDE.
-
IDE and by-products. The controller is installed on the disk and
the interface you/he/she is reduced to a door, usually integrated
in the mother card. The connection happens with a solo cable
to 40 poles. The physical limit is of two unities for door (Master
and Slavics).
-
SCSI. The connection with the PC happens through a specific controller,
that allows a macro discussion you command, type bus, on the
cable of connection that allows to have up to 7 unities in parallel.
The cable, single, you/he/she can be to 50 poles, for the type
of interview to 8 bit or 68 poles for that to 16bit (SCSI Wide).
More controllers can cohabit on the same sitema increasing the
number of unity installabili.
GO UP
Firmware
From how much I dictate before, it
is by now a datum of fact the presence of one or also more processors
on the electronic cards climbed on in the hard disks. Obviously,
if a processor is present, there will be also a chip you contain
the software for the operation of the complex. This takes the name
of firmware.
Substantially it deals with a mini operating system that contains
the routineses of control of the hardware, the procedures of coding
and the control of the interface toward the principal system.
You/he/she is mentioned to more processors as some functions, the
interface, is submitted for instance to a processor, while the complex
operations of writing and reading are developed by another unity.
The motors are also checked from integrated endowed with an elevated
"intelligena" and autonomy.
The firmware can be contained in the processors, that have a programmable
inside area as PROM or EPROM, or in an external chip also riprogrammabile
(EEPROM, FLASH). You are not thought, however, with this to the
possibility of upgrade of the "BIOS" of the disks; different motivations,
among which the structure of the hardware and also the necessity
to protect the safety of the content from the reverse engeneering
they do him that this possibility is not one of the offers of the
builders of disks.
For engraved, cases of conflict are verified between the firmwares
of disks and those of the BIOSs of the mother cards; in these cases,
gives the impossibility to change the firmware on the disks, the
solution it is in the upgrade of the BIOSs of the mainboards. If
this is not possible, other doesn't stay from unfortunately to do
whether to replace one of the two components.
Other problems of the firmware can concern the incompatibility of
a drive with another of different brand or problems with the managements
DMA (Bus Mastering).
GO UP
The electronic circuit of inside management of the signals
from and for the heads.
The traces the magnetic ones on the
disks have dimensions a lot of redoubts and they have to be glances
or written in brief times to get some elevated performances. The
operation to check the flow of the tide in the heads is realized,
in the modern hard disk, with special circuits devoted to this purpose,
usually places inside the disk, how much nearer possible to the
heads.
This is made necessary by the necessity to reduce to the least one
the length of the conductors that you/they bring the signals from
and for the heads themselves.
This integrated is found positioned on a flexible printed circuit
that also constitutes the connection among the arm it brings heads
and the connector that it brings the signals to the principal printed
circuit.
 In the image you/he/she
can clearly be seen the big chip of control of the heads climbed
on on the flexible printed circuit, that has been drawn out from
the inside of the box of the hard disk. On the flexible circuit
other necessary components are noticed to also the operation or
resistances and assemblage condensers superficial (the dark rettangolinis
with the metallic extremities). Á. the left extremity there is the
connector for the connection with the printed circuit of the controller.
On the right side the arm is seen it brings heads with three groups
of 2 heads (3 disks).
Digital information are constituted by sequences of 1 and 0 that the electronic
circuits turn into levels of tension (1 = tall level, 0 = low level).
On the magnetic support information are stored as I impelled magnetic.
The circuit of the heads has, therefore, the purpose to convert
the logical levels in magnetic levels and vice versa.
Magnetic information consist in dwarfish areas of the surface of
the disk in which a magnetic field is preserved. We can think about
dwarfish magnets, wide some about ten milionesimi of millimeter.
You think to a normal magnet: it has two poles, said north and south
and the magnetic energy (the magnetic field) it flows among the
two poles; magnetic information are therefore of the dwarfish magnets.
How do you/they do to remain on the disk? For it persuades her/it
of some material ones to preserve a part of the magnetic field that
has been applied him.
You try to take with a magnet a platforms you tap or nails: many
of them, removed once the magnet, they will stay magnetized and
they will be able to them it turns to lift other pins or nails.
You now suffered to replace the magnet with an electromagnet (the
head) and you will have the principle on which the magnetic writing
of the disk founds him (and of the floppy, ribbons, the cottages
of the mangianastri, etc.).
And it is note thinking about the cassettes audio that we can have
a mink it completes of a procedure of magnetic scrittura/lettura,
analogous to that of the hard disk. The comparison is suitable,
in first appeal; all it takes is changing some expressions as in
the chart that follows:
Recorder audio (incision) |
Magnetic disk (writing) |
The voice (vibration of the air) you/he/she is turned into electric signal
through the microphone |
The software turns the data into binary logical signals |
the circuit of recording elaborates the signal and sends him/it to the
head |
the circuit of control of the disk elaborates the signal and sends
him/it to the controller of the heads |
the head converts the electric signal in a magnetic field and him "it
engraves" on the ribbon, in the form of magnetic field |
the head converts the electric signal in a
magnetic field and him "it engraves" on the ribbon,
in the form of magnetic field |
Once "written", the magnetic material of the ribbon (or of the disk) it
will preserve for long time the magnetic fields that constitute
the recording; it can be "read" with an inverse procedure:
Recorder audio (reading) |
Magnetic disk (reading) |
the ribbon flows in front of the head and the variations of the engraved
magnetic field they are turned into variations of electric
level |
the surface of the disk rotates in front of the head and the variations
of the engraved magnetic field they are turned into
variations of electric level |
the circuit of reading elaborates the received signal and amplifies him/it
to a suitable level |
the circuit of reading elaborates the received signal and turns him/it
into digital signals |
the loudspeaker turns the electric signal into audible sound (as I enliven
mechanic of the air) |
the software allows to extract from these digital signals the stored
information |
Clear, no?
In effects the example is rather ok; to be precise, however, is
necessary to add that the ribbon audio contains a magnetic field
"continuous", varying as value according to the modulation of the
sound that has to preserve. On the disk, instead, some small defined
magnetic fields are engraved.
The inspector of the heads has fixed on a flexible printed circuit: it
deals with a particular type of printed circuit realized with material
extremely flexible and resistant and in degree of million and million
of times to be flexed without getting deformed himself/herself/themselves
or to break himself/herself/themselves following the movement of
the heads.
The signals pre treated by the circuit of control of the heads is
transferred then to the principal circuit where you/they are elaborated.
Different methods exist for the parceling, the writing and the reading
of the data. The actual tendency is that to always place a greater
number of data to parity of magnetic surface; this is gotten with
to progress of the constructive technologies and of the materials
and with an increase of the complexity of the circuits and the systems
of reading and writing, with the introduction of particular heads,
of processors and techniques DSP (Digital Signal Processing) a great
deal complex.
GO UP
Coding of the signal.
You/he/she could be brought to think
that the polarities north and south of the magnet can be employed
as 1 and 0 platforms; in effects the thing is not so simple because
through the heads it is possible to read with simplicity not the
polarity of the field but its variation, for instance that coop
happening a transition from a field directed north-south to one
directed south-north.
The connected problems with the maintenance of microscopic magnetic
fields, tightly approached one to the other, with opposite polarity
it also asks for particularly complex technologies, in how much
smaller they are the surfaces interested by the fields (more fields
for unity of surface = more information immagazzinabili on the disk)
it meets him with the necessity of magnetic materials of quality
and purity it excelled and with the construction of extremely small
heads, but in degree to produce very intense fields (for the magnetization
of the surface in writing) and to notice with safety the micro variations
among the fields (during the reading).
Besides this it is always in demand an elevated speed of lettura/scrittura
for the greater application of performance; the increase of the
transfer installments (quantity of transferred data) it is gotten
with a more rapid rotation of the disk so that to bring under the
heads a greater surface in the unity of time.
This requires that the data are not simply thrown them, but are
encoded (encoding) according to systems and particular algorithms.
Besides the codings have the purpose to crowd how much more possible
data on the useful surface.
In this sense, an example of "coding" that you/he/she can approximately
make the idea of one of the finalities of these systems it is the
known compression of the data, perlappunto with "codings" as ZIP,
ARJ, etc.
Of thing he treats, in substance? If we think that a whole organized
of data (file) it is composed from a certain number of bytes, to
es. 100, written on a disk it will occupy a certain surface. If,
with some system, I can compress these information so that the spread
out content is available in 50 byteses, I will have the possibility
to store on the surface of the disk the double one of information
(and therefore I will have in a certain sense increased it persuades
her/it some disk).
The methods of coding however I have to also deal himself/herself/themselves
with the physical problems of the writing and reading of data on
a rotating support to high speed; the everything, obviously, maintaining
a safety standard and reliability how much taller possible. For
instance, a notable problem is that to identify where a field finishes
and starts a following of it (to remember himself/herself/themselves
that we are speaking of fields of infinitesimal dimensions that
rotate under the heads to 5400, 7200 or 10000 turns to the minute!).
For instance, wanting to store a lace of 1 logical we won't be able
certain to deposit an equivalent quantity of magnetic fields of
the same polarity or intensity; it would be extremely difficult
to identify where a field finishes and starts the following one...
A possible solution is that to tie the writing to a clock: every
information is spatially written and temporally in narrow bond with
a signal of synchronism that helps to identify her/it during the
reading, that will be syncronized to the same clock.
More complex other methods of coding have been developed for being
able to always store in the disk a greater quantity of data. In
substance, to conclude, the technologies hardwares develop him with
the purpose to make for unity of surface possible always greater
densities of magnetic fields, while the methods of coding try to
crowd a greater quantity of information in the fields, a writing
and reading making at the same time possible reliable.
In the following paragraphs we see some of the more diffused methods
of coding.
Frequency Modulation (MFM)
One of the first systems adopted for codifying the digital signals
to save on a magnetic mean have been that of the modulation of frequency
(FM Frequency Modulation). The concept is similar to that of the
issues I remove FM (Modulation of Frequency). In this method a datum
of value 0 are written as two consecutive magnetic fields of opposite
polarity, while a 1 are constituted by two fields of the same polarity.
The signal of writing is syncronized with a clock: the first magnetic
field, correspondent to the first hit of the clock, constitutes
him/it "start" of the bit, while the following one is the value.
The definition of modulation of frequency derives from the fact
that the reading is "in movement." If we point out with N the inverted
field and with P that not inverted, the 0 logical it is representable
with NP, while the 1 with NN; a byte composed of 8 bit of value
0 will be NPNPNPNPNPNPNPNP, while a byte of 1 will be NNNNNNNNNNNNNNNN
being tied up to a clock; he/she is seen how the 1 representation
frequency is double of that of the 0.
Currently FM, broadly used in the first systems of magnetic memory,
it is not more proper for the actual solutions. The principal limit
is in the necessity of two magnetic fields to define a bit.
Modified Frequency Modulation (MFM)
MF, with to appear some first hard dishes rotating disk, you/he/she
is replaced by MFM or modified frequency modulation (Modulation
of Modified Frequency) that it reduces the number of the necessary
magnetic fields to define a bit inserting only an inverse field
in presence of two consecutive zeros. This way doing can be reached
also the doubling of the it persuades some disk.
The method MFM has been employed around first hard disk (interface
St-506) and it is today still the system of coding for the floppy
disk
Run Length Limited (RLL)
Already with the series of hard disk with interface St-506 or SCSI
of it persuades superior to the 40MB the method RLL is introduced
(Run Length Limited or limited operational length). In effects it
doesn't deal with a single method, but of one "family" of methods
of coding sophisticated, thesis to overcome the limits imposed by
the simple modulation codings of frequency.
RLL operates not on single bit, but on groups of bit, employing
both clock and to get packets of data of great compactness that
allow a writing and reading more efficient and sure.
RLL has two principal parameters: run length and run limit, from
which it derives the name. Run length is the least space (tempo/superficie)
among two inverse magnetic fields, while run limit is the maximum
allowed. The tempo/spazio among two inverse fields has to be rather
small, otherwise the head of reading risks to lose the synchronism
with the clock. Contemporarily the heads have to become more and
more small to allow the precise writing of the fields. A beautiful
match!
The parameters of RLL are express in the form "run length, run limit
RLL"; for instance a common type can be 1,7 RLLs.
The sophisticated codings of the type RLLs ask more and more for
controller complexes; the electronic cards on board some disks include
by now microprocessors, quartzes, memories, encircled devoted.
Partial Response, Maximum Likelihood (PRML)
Having microprocessors on board of the disks can be done a lot.
The systems of writing thin to you now approve they found him on
the survey "to the flight", through the head, of the variations
among the direct and inverse magnetic fields, analyzing the variations
in relationship with a clock.
With to increase some density of
the data and the speed of rotation of the dishes a limit it is reached
beyond which the analysis of the variations of field becomes problematic
and the possibility of error is amplified. The signal from the heads
becomes less and less clean, less and less digital and never analogical,
making difficult to the classical digital circuits the definition
of the areas of the elementary magnetic fields that you/they constitute
the core of the information.
The builders are to the continuous search of new solutions, mainly
based on the digital analysis of the signals, through special says
processors DSP (Digital Signal Processor) in degree to operate to
high speed on the information coming from the heads.
Quantum has for instance developed a called system Partial Response,
Maximum Likelihood (PRML) that it employs complex techniques hardware
and software. It is not looked for more than to identify the single
fields, but, through the DSPs and opportune algorithms, blocks of
analogical data transform him read by the heads (partial response)
to determine the correct sequence of bit that has greater probabilities
to be that that has produced the writing of that specific sequence
of fields (maximum likelihood).
Every builder develops technologies to hoc and the development you/he/she
is still in action.
GO UP
Transfer Rate
The transfer installments or rate
of transfer it points out the quantity of data that you/they can
be trsferita from and for the disk. It is obvious that a disk with
a transfer installments stop it will have best performances of one
that exhibit a low value.
Its value depends on many factors; it increases with to increase
some speed of rotation, with the improvement of the methods of coding
and correction of the errors, the formalities of operation of the
interface, etc..
Specific tests exist for determining this factor, but you/they must
be taken with the opportune cautions. We see because.
You/he/she has happened that a test on the disk effected with a
commercial benchmark has given a good result on the new disk and
then, repeated after a certain time, has given cheaper results.
How come?
The problem is in the fact that the transfer of the data from the
dishes is not constant but it depends on the position of the heads
in comparison to the edge of the disk. The peripheral zones have
a greater extension and you/they can contain greater density with
a speed of greater transfer of those more inside.
This explains the mystery of which above: the first test is effected
on an almost empty disk and you/they is employed the external traces,
more performantis; the following tests, done when the disk is fuller,
they move the area of the test to the inside zones, less performanti
() there can be also a relationship 1 to 2!).
The solution is the employment of test that keeps in mind this problem
(and in general the to take the results of the tests "cum wheat
salis"...).
For engraved, you/he/she must be remembered that also the fragmentation
of the files is a cause of the low performances of the tests; before
the tests it is opportune to effect a deframmentazione (defrag).
Also in the common use it is opportune to sometimes remember himself/herself/themselves
deframmentare the disk, both to improve the performances, both to
reduce the possibilities of error.
GO UP
Write Precompensation
Write precompensation is reported
to the necessity to vary the parameters of the tide in the heads
during the writing on the disks. The old disks used the same number
of sectors for trace, independently from the fact that it dealt
with external or inside traces. It is evident that on the inside
traces, of smaller circumference, the sectors have to be of smaller
dimension of those on the most external traces, of greater circumference.
In substance, the density of magnetic fields is not uniform on the
whole disk. This door to the necessity to compensate somehow this
difference that, over a certain trace, the rilettura would make
very difficult; therefore, one of the present parameters in the
charts of the old disks are the number of trace from which to depart
with the compensation (what you/he/she is managed by the system
of control really of the disk).
This parameter is not more necessary for the modern disks that use
different systems of writing and reading and endowed with very evolved
controller.
GO UP
Interleaving
Another term that applies when it
speaks of disks it is the word interleaving (seen translated in
Italian with the horrible term interleaving) it points out it needs
her/it the presence of one "space" vital among the sectors.
The sectors of a disk are numbered in a logical sequence so that
to be able to be addressed without problems during the operations
of writing and reading. However it is not said that a logical sector
1 are physically set before the sector 2; this depends on the systems
of coding, of management of the disk, etc.. To the practical action
a logical sector and the following one, if they were found adjoining,
could create problems in the reading! Because this? because it needs
to remember that the disk turns to great speed under the head and
at the end of the reading of a sector the system it asks for a certain
time to organize the read data and this could be such to be made
to escape the following sector and to ask for a new turn of the
disk for the reading. If, instead, the sectors are not adjoining
to es. the sequence is 1 - 3 - 2, read the sector 1, the whole time
of the passage of the sector it stays 3 before beginning the 2 reading,
leaving so the whole time to the circuits to be able to proceed
to the recovery of the data from the sector 2 without asking for
another turn of the disk.
Obviously the best case is an interleave of 1, that is all the logical
and physical sectors are corresponding; but it is possible solo
implying one it persuades of such reading not to have dead times
at the end of a sector (to es. exploiting a cache). Otherwise the
performances are greatly reduced.
The standard of the old disks was typically of 17 sectors for trace.
With an interleave of 1:1, the sequence of the physical and logical
sectors corresponds (or 1-2-3-etc.). With an interleave of 2:1 the
sequence typical of logical numeration of the physical sectors becomes
1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7, 16, 8, 17, 9. In this
case a non consecutive sector is inserted in the sequence and the
logical disposition it is such for which, if the controller asks
for the time of passage of a sector to complete the operations on
the precedent, the performance on the whole trace it is the possible
motto.
According to as the alone disk a specific interleave is built it
is able to furnish the maximum performances and it is not said is
1:1. The common values went from 1:1 to 5:1. For this the utilitieses
of management of the disks, in the phase of low-level formatting,
they foresaw a test (middle analysis) to determine what interleave
was the most proper.
In the actual disks the parameter of interleave is not more necessary
neither accessible from the outside; all the controls and the parameters
of formatting of the disk are planned by the builder and it is not
anticipated that the consumer manipulates them in some way.
The fact that some mother cards preserve in the BIOS the utilitieses of
low-level management of the disk or toolses are available with analogous
functions it doesn't authorize to use them; to remember that it
is not possible to format low-level the disks IDE if not with specific
tools of the builder, well rarely available to the consumers.
GO UP
The errors!
But is a disk wrong? Certainly, only
that is organized so that the system I/you/he/she don't realize
of it.
The causes of error are manifold; modern technologies have pushed
to the extreme the materials and the high density of writing, her
not perfect uniformity of the magnetic materials, the high speed
of rotation and the intense flow of data can give errors origin
in the reading. Electromagnetic troubles, current induced, thermal
problems are other causes of error.
As an error on the data saved on the disk would be unacceptable, all the
builders are hocked to put all the possible solutions to prevent
this problem into effect. Obviously we are not talking of due errors
to crash of the motor or the heads or to damages to the electronics
of control: these put out use the disk partly or totally and for
their prevention opportune mechanisms have been implemented (SMART).
The basic system for the revelation of the errors and the correction in
transparent way is Etc (Error Correcting Tails). Similar to that
implemented in the memories RAM, with analogous functions, it consists
in algorithms, to es. the Reed-Solomon, that allow to the electronics
of control to correct situations of error owed to the wrong reading
of one or more bit. Usually the algorithms found him on the redundance
of the information and they foresee sophisticated routines software.
A sector typically contains 512 byteses
or 4,096 bit; to these other bit are added devoted etc to the. The
quantity depends from the used algorithm and from the planning of
the system; it is had to mediate between the safety of the correction
and the reduction of the space and the performances. When a sector
is written, the relative number of bit is written of also Etc; when
the sector is reread, the algorithm combines data and Etc and, verified
an error, it corrects him/it in the limits imposed by the planner.
This operation, as says, it is transparent entirely for the consumer,
even if in some disks of the most recent families, the data of intervention
of the correction are monitorati from the circuit of control both
to activate mechanisms of inside backup (substitution of bad sector
with others backup) both to signal the possible serious spoiled
future of the disk (SMART).
If the systems of correction of the error are not able to accomplishedly
intervene, then the breakdown is brought to the consumer.
A possible sequence of the interventions is:
-
Revelation of the error: to the data read in the sector the procedure
you/he/she is applied Etc and, if errors are not verified, the
data are envoys to the interface to have made of the system
available.
-
Correction of the error: Etc the algorithm corrects the error in
reading using the redundant information. A correct error to
this level is not considered really an error.
-
Repetition of the reading: if Etc the system has not been able
to correct the error because too wide for its possibilities,
the following footstep is a new attempt of reading of the sector.
This can automatically be served as the circuit of control of
the disk. An error is often caused from a problem of the magnetic
field or from other non repetitive causes and the rilettura
of the magnetic zone it allows the correction of the error.
In this case it speaks of data "recovered"o of correct error
after a new attempt.
-
Advanced procedures of correction of the error: many builders implement
procedures of corrections that involve more sophisticated algorithms
and, usually, in degree to correct the error. Because, then,
not to directly use them? Because the complexity of the procedure
would bring to a deceleration of the transfer of the data that
is acceptable for an irregular error, but it would penalize
the performances in the current operation. At times the procedures
involve also the hardware, replacing, for instance, the damaged
sectors with others kept as it reserves.
-
Non corrigible error: if none of the preceding procedures is able
to correct the error the driver it will signal the breakdown
to the system.
Map of the bad sector
The structure of the old hard disk in which the correspondence among
physical and logical parameters was very narrow and that they didn't
have the sophisticated actual technologies, one of the typical elements
were the presence of a list of bad sector. The builder usually furnished
the disk with a sheet of testing comprendente the list of the sectors
found defective to the test, or of the areas of the dishes in which
defects of metallizzazione made the writing of the data insecure;
the consumer would have inserted then these paramentris in the oportunes
charts during the low-level formatting del'unitá, to allow the system
to exclude the defective areas from the assignment of addresses
him logical.
To the actual state, also this function doesn't belong to the necessities
of modern hard disk that is structured in different way, with logical
assignments of the established physical surfaces during the productive
trial and not accessible or modifiable from the consumer.
GO UP
|
