Track on CD scanword. Select and copy tracks to your hard drive. Required for work

My previous article was devoted to the internal structure of the chip from Nvidia, and, perhaps, the internal structure of any modern processor. In this article we will move on to information storage media, and I will tell you what CD and HDD drives are at the micro level.

CD

Let's start with the CD. Our test subject is a simple CD-R from Verbatim. A regular disc with recorded (or rather, printed) information consists of 3 main layers. Layer A is a polycarbonate disk that is responsible for several functions at once. The first is the disk base, which can withstand the enormous rotation speeds inside the drive.


So in general terms you can imagine the structure of a CD disk

The polycarbonate disk, as it turns out, is additionally coated with a special varnish that protects against lung mechanical damage outer surface of the disk.


The varnish layer is highlighted in red, with polycarbonate “beginning” underneath it


Under the beam of an electron microscope, the layer of protective varnish does not feel very good

Secondly, it is on polycarbonate, in the literal sense of the word, that information from the matrix is ​​printed - be it a film, music or programs. As Wiki tells us, the polycarbonate base is 1.2 mm thick and weighs only 15-20 grams.

Naturally, polycarbonate and varnish are transparent to laser radiation, so the “printed” information must be made “visible” for the laser, for which the surface is coated with a thin layer of aluminum (layer B). It is worth noting that CD-ROM with “printed” information, CD-R and CD-RW have minor differences. In the last two cases, an intermediate layer is added between polycarbonate and aluminum, which can change its properties under the influence of laser radiation of a certain wavelength, and empty tracks are printed on the polycarbonate. These can be either dyes in the case of CD-Rs (something similar to photoresist) or metal alloys in the case of CD-RWs. That is why it is not recommended to expose rewritable discs to direct sunlight and overheating, which can also cause changes in optical properties.

Let's compare the disk and the aluminum layer torn from it. It can be seen that on the polycarbonate there are “grooves” (pits), and on the aluminum layer, on the contrary, there are elevations that completely correspond to the grooves:


Familiar depressions on the surface of polycarbonate (AFM image)


The “opposite” pits are visible on the protective aluminum layer: not grooves, but protrusions (AFM image)

Next, the resulting “pie” is covered with a special protective layer C, whose main duty is to protect the “delicate” aluminum reflective layer. Then you can stick something on this layer, write with a marker, apply special additional layers for printing, etc. and so on.

This video presents all technological stages CD production:

A recording on a CD is similar to a recording on a vinyl record, i.e. The information path goes in a spiral. It originates at the center of the disk and ends at the outer edge. But right in the middle of the disk, empty sections and tracks with recorded information are “joined”:




There was a recording, but there wasn’t. Comparison of empty tracks and tracks with recorded information (SEM micrographs)

There are no fundamental differences at the micro level between CDs and DVDs and, probably, Blu-Ray. Unless the pitas will be smaller in size. In our case, the dimensions of 1 minimum recess are 330 nm in width and 680 nm in length, while the distance between the tracks is ~930 nm.

N.B. If you have a scratched CD that cannot be read in any drive, try polishing it. Almost any clear polish will work for this. It will fill in the indentations that interfere with reading the information, and you will at least be able to copy the information from the disk.

How, after all, sometimes a layer of aluminum bends bizarrely (practically a work of art - black and white):


Black and white stripes of our life. CD (SEM micrograph)

And finally, a couple more images of the CD obtained using an optical microscope:


Optical microscopy: left - aluminum reflective layer, right - Al layer (lighter area) on a polycarbonate disk (darker area)

HDD

Let's now move on to the hard drive. It has always been a mystery to me, since the days of floppy disks and VHS, how magnetic memory works?! Before writing the article, I tried to find at least some video and media materials that would demonstrate, as in the previous video, the main stages of hard drive production, and Vicki was unpleasantly pleased: “Both planes of the plates, like a tape, are covered with the finest ferromagnetic dust - oxides of iron, manganese and other metals. The exact composition and application technology are a trade secret." I had to come to terms with it and not seek the truth from HDD manufacturers (unless Seagate slightly revealed its secrets), especially since with the advent of the SSD era, competition in the market intensified even more.

The plates themselves are made of non-magnetic metal alloys. The basis of these alloys is aluminum and magnesium, as the lightest structural materials. Next, a thin, again according to Wiki, 10-20 nm layer of magnetic material is applied to them - here, perhaps, the word nanocrystalline would be appropriate - material, which is then covered with a small layer of carbon for protection. Since the disk is NoName, and it is made using the ancient technology of parallel recording of information, I will allow myself to give here the composition of the material according to EDX data (X-ray spectral microanalysis): Co – 1.1 atomic %, Y – 1.53 at. %, Cr – 2.38 at. %, Ni – 45.81 at. %. Carbon content 36.54%. Si and P came from somewhere, the content of which is 0.46 at. % and 12.25 at. %, respectively. The origin of silicon - apparently, remained in trace quantities on the surface after the work of the microtome and my polishing, and phosphorus - simply stained the sample.
Honestly, I tried to find a layer of magnetic material with a thickness of “10-20 nm”, but without success. Based on what I saw, the surface layer is approximately 12 micrometers thick:


That very “thin” layer that stores information in our hard drives

Of course, you can correct me in the comments, but:
1. the disk is quite old (i.e. its production date dates back to the beginning of the last decade);
2. the features of EDX are such that the signal output depth ranges from 1 to 10 microns;
Thus, it seems to me that these 12 micrometers are a magnetic layer, which is covered on top with a thin layer of carbon (50-100 nm), which may not be visible in the section.

The surface of the disk itself is very, very smooth, the height difference is within 10 nm, which is comparable to the surface roughness of monocrystalline silicon. And here are the images in phase contrast mode, which correspond to the distribution of magnetic domains on the surface, i.e. we actually see individual bits of information:


AFM images of the hard drive surface. On the right are phase contrast images.

A little about phase contrast: first, the AFM microscope needle “feels” the relief, then, knowing the relief and repeating its shape, the needle makes a second pass at a distance of 100 nm from the sample in order to “muffle” the action of van der Waals forces and “highlight” the action of magnetic forces strength You can watch a flash drive about how this happens.

By the way, have you noticed that individual magnetic domains are extended along the plane of the disk and are parallel to it?! Let me say a few words about recording methods. On this moment disks with a perpendicular method of recording information (i.e. those in which the magnetic domains are oriented perpendicular to the plane of the disk), which appeared in 2005, have almost completely replaced disks with parallel recording. The advantage of perpendicular recording is obvious - the recording density is higher, but there is one subtle point in connection with Wiki’s data on the thickness of the magnetic layer. This nuance is called the superparamagnetic limit. Those. There is a certain critical particle size, after which the ferromagnet, already at room temperature, transforms into a paramagnetic state. Those. There is enough thermal energy to rotate and reorient such a small magnet. In the case of magnetic recording, they often do the following: they make one of the “magnet” sizes larger than the other two (this is clearly visible in the picture with the distribution of magnetic domains), then in this larger direction the magnetic moment is preserved. So, if in the case of parallel recording I can still believe that the magnetic layer is tens of nanometers with the size of 1 bit being several micrometers, then in the case of perpendicular recording this simply cannot be. The thickness of such a magnetized region at minimum sizes in the plane of the disk, it simply must be at least a few micrometers. So, perhaps Vicky is cheating a little. Or they apply a magnet in the form of nanoparticles with a diameter of 10-20 nm, and only then in some “cunning” way they divide the disk into areas that are responsible for storing information. Unfortunately, I did not fully satisfy my curiosity and answer questions about magnetic recording of information, maybe someone can help?!


Comparison of parallel and perpendicular methods of recording information on hard drives

Maybe someone will like the video in English from Seagate:

The latest on how the cost of 1 Mb has changed since 1995 HDD drive and how many discs were released:

As promised, I’m posting a video about how the shooting was carried out using various devices (don’t forget to read the description of the video on YouTube and leave your comments). For statistics: filming took 4 days (although everything could have been done in 2), the duration of the video that was edited was about 3 hours, and the result was a 15-minute video. I hope there will be English subtitles for this video soon.

P.S.: This article was published on the eve of the Science Festival, which will be held in Moscow from October 7 to October 9, 2011 (in reality Free access will only be on October 8 and 9), and I would like to invite everyone to visit our exhibition “The Beauty of Materials,” which will be held on the second floor of the Fundamental Library on the territory of Moscow State University.

P.P.S.: Together with Anton Voitsekhovsky, we are preparing several video notes on how some biological objects work (a rose, for example, looks simply gorgeous). I think that they won’t appear on Habré (you must admit, it’s difficult to link a microphotograph of a razor or a match head to IT), but as soon as the videos are ready, they will immediately appear on my channel on youtube and rutube, and definitely on the Nanometer.ru website.

Opening the Nvidia 8600M GT chip, a more detailed article is given here:

Track - this is one “ring” of data on one side of the disk. A recording track on a disc is too large to be used as a storage unit. In many drives, its capacity exceeds 100 thousand bytes, and allocating such a block to store a small file is extremely wasteful. Therefore, the tracks on the disk are divided into numbered sections called sectors .

The number of sectors may vary depending on the track density and drive type. For example, a floppy disk track can contain from 8 to 36 sectors, and a hard drive- from 380 to 700. Sectors created using standard programs formatting have a capacity of 512 bytes, but it is possible that this value will change in the future. One thing to note important fact: for compatibility with older BIOS, regardless of real quantity sectors per track, the device must translate to the 63 sectors per track mode adopted in CHS addressing.

The numbering of sectors on a track starts from one, in contrast to heads and cylinders, which are counted from zero. For example, a 1.44 MB floppy disk contains 80 cylinders, numbered 0 to 79, has two heads (numbered 0 and 1) in the drive, and each cylinder track is divided into 18 sectors (1–18).

When formatting a disk, additional areas are created at the beginning and end of each sector to record their numbers, as well as other service information, thanks to which the controller identifies the beginning and end of the sector. This allows you to distinguish between unformatted and formatted disk capacity. After formatting, the disk capacity decreases, and you have to put up with this, since in order to ensure normal operation of the drive, some space on the disk must be reserved for service information. It is worth noting, however, that new disks use formatting without an identifier, i.e. the beginning and end of each sector are not marked. This allows you to use a little more space to store actual data.

At the beginning of each sector, its header (or prefix) is written, which determines the beginning and number of the sector, and at the end - the conclusion (or suffix), which contains the checksum necessary to check the integrity of the data. In the above-mentioned addressing system without identifiers, the beginning and end of each of the sectors is determined based on the pulses of the clock generator.

In addition to the indicated areas of service information, each sector contains a data area with a capacity of 512 bytes. In low-level (physical) formatting, all bytes of data are assigned a value, such as F6h. Electronic circuits drives have great difficulty encoding and decoding certain patterns because these patterns are only used in the drive manufacturer's testing during the initial formatting process. Using special test patterns, you can identify errors that are not detected using regular data patterns.

Note!

Low-level formatting is discussed next. Do not confuse this with high level formatting, which is done with FORMAT programs in DOS and Windows.

Sector headers and suffixes are independent of the operating and file systems, as well as from files stored on the hard drive. In addition to these elements, there are many spaces within sectors, between sectors on each track, and between tracks, but none of these spaces can be used to write data. Gaps are created during low-level (physical) formatting, which deletes all written data. On a hard disk, the gaps perform exactly the same functions as on a tape cassette, where they are used to separate music recordings. Leading, trailing, and intervening spaces are precisely the space that makes the difference between formatted and non-formatted disk capacity. For example, the capacity of a 4-megabyte floppy disk (3.5-inch) after formatting is “reduced” to 2.88 MB (format capacity). A 2 MB floppy disk (before formatting) has a format capacity of 1.44 MB. The Seagate ST-4038 hard drive, which has an unformatted capacity of 38 MB, is “reduced” to 32 MB (formatted capacity) after formatting.

Low level modern formatting hard drives ATA/IDE and SCSI are done at the factory, so the manufacturer only specifies the format capacity of the disk. However, almost all disks have some reserved space to manage the data that will be written to the disk. As you can see, it is not entirely correct to say that the size of any sector is 512 bytes. In fact, each sector can store 512 bytes of data, but the data area is only part of the sector. Each sector on a disk typically takes up 571 bytes, of which only 512 bytes are allocated for data. IN various drives The space allocated for headers and suffixes varies, but generally a sector is 571 bytes in size. As already mentioned, many modern disks use a partitioning scheme without sector header identifiers, which frees up additional space for data.

For clarity, imagine that the sectors are pages in a book. Each page contains text, but it does not fill the entire space of the page, since it has margins (top, bottom, right and left). The margins contain service information, such as chapter names (on a disk this corresponds to track and cylinder numbers) and page numbers (which correspond to sector numbers). Areas on a disk, similar to margins on a page, are created when the disk is formatted; At the same time, service information is also recorded in them. Additionally, during disk formatting, the data areas of each sector are filled with dummy values. After formatting the disk, you can write information to the data area as usual. The information contained in the sector headers and conclusions does not change during normal data write operations. You can change it only by reformatting the disk.

The table shows the track and sector format as an example standard hard disk having 17 sectors per track. The table shows that the “useful” volume of the track is approximately 15% less than possible.

These losses are typical for most drives, but they may vary for different models. Below we analyze in detail the data presented in table. 9.2. The post-index interval is needed so that when the head moves to a new track, the transient processes (installation) end before it reaches the first sector. In this case, you can start reading it immediately, without waiting for the disk to complete an additional revolution.

The post-index interval does not always provide enough time to move the head. In this case, the drive gains additional time by shifting sectors on different tracks, which delays the appearance of the first sector. In other words, the low-level formatting process causes the sector numbering to shift, causing sectors on adjacent tracks that have the same number to be offset from each other. For example, sector 9 of one track is adjacent to sector 8 of the next track, which in turn is adjacent to sector 7 of the next track, and so on. The optimal displacement value is determined by the ratio of the disk rotation speed and the radial speed of the head.

Note!

Previously, the head offset parameter was manually set by the user when low level formatting. Today, such formatting is done in an industrial environment, and these parameters cannot be changed.

The sector identifier (ID) consists of cylinder, head and sector number recording fields, as well as a CRC control field to verify the accuracy of reading the ID information.

Most controllers use the seventh bit of the head number field to mark bad sectors during low-level formatting or surface analysis. However, this method is not standard, and in some devices bad sectors are marked differently. But, as a rule, a mark is made in one of the sector identifier fields. The write enable interval immediately follows the CRC bytes; it ensures that the information in the next data area is written correctly. In addition, it serves to complete the analysis checksum(CRC) sector identifier.

The data field can store 512 bytes of information. Behind it is another CRC field to check whether the data was written correctly. In most drives, the size of this field is 2 bytes, but some controllers can work with longer Error Correction Code (ECC) fields. The error correction code bytes written in this field allow some errors to be detected and corrected when read. The effectiveness of this operation depends on the selected correction method and the characteristics of the controller. The write-off interval allows the ECC (CRC) byte analysis to be fully completed.

The interval between records is necessary in order to insure the data of the next sector from accidental erasure when writing to the previous sector. This can happen if, during formatting, the disk was rotated at a speed slightly lower than during subsequent write operations. In this case, the sector will naturally be a little longer each time. Therefore, so that it does not go beyond the boundaries set during formatting, they are slightly “stretched” by introducing the mentioned interval. Its actual size depends on the difference in disk rotation speed when formatting a track and each time the data is updated.

The pre-index interval is necessary to compensate for the uneven rotation of the disk along the entire track. The size of this interval depends on the possible values ​​of the disk rotation speed and the synchronization signal during formatting and recording.

The information recorded in the sector header is of great importance because it contains information about the cylinder, head and sector numbers. All of this information (except for the data field, CRC bytes, and write-off interval) is written to the disk only during low-level formatting.

After the disk information is received and written to database, a dialog will appear in the working window displaying the disc tracks with their names and other information (Fig. 2.41). This dialog is intended for selecting tracks to save, as well as for setting parameters for this saving.

At the top of the dialog a list of all audio tracks on the disc is shown.

You can mark the required tracks for saving, and below the field you will see the total time and volume of the selected tracks.

To select a track, click on it with the mouse. To select multiple tracks, hold down the key while selecting. A group of tracks can be selected by holding down a key and clicking on the first and last tracks in the group. Rice. 2.41. Selecting tracks You can listen to the selected tracks using a simple player. Below the list of tracks is a slider that displays the playback position of the current track, and even lower are the playback control buttons. By pressing the buttons, you can play a track, stop playback, go to the next or previous track, and so on. Drop-down list Format(Format) allows you to select one of the formats for storing audio data on disk. To further record music discs, select the PCM Wave element from this list. In the input field

Path Format(Path) you should enter the name of the folder where the selected tracks will be saved. And in the input field Method for creating filenames(File name creation method) specifies the name of the audio track to be saved. You can choose several options for creating a name. To do this, open the list(File name creation method), and the name selection dialog will appear (Fig. 2.42). By setting the switch, you can choose a manual naming method, where you assign a name to each track. If the disk information is known to the program, then the middle position of the switch becomes available. In this mode, the track name is formed as the artist name and song name separated by a dash character. If the disk is not recognized, then the name is formed as a word"Track" and track number. The lower position of the switch is called Personal (User Defined) and allows you to create names as you wish. In this case, you can use any characters, as well as special character sets. So,%A indicates the name of the artist,%N - track number, to confirm your selection.

Rice. 2.42. Name creation options

Pressing the button Options(Options), you will open additional controls. If you hear distortion when listening to saved tracks from a CD, try copying the track again by checking the Jitter correction(Jitter Correction). Additionally, if you want to remove pauses between tracks, you should check the box Delete pause(Remove Silence).

When copying tracks, the program can automatically create Playlist(Playlist). The list contains the tracks in the order in which they were copied. Using such a list is useful if you want to create discs containing music encoded in the MP3 format. Many home and portable devices use a playlist in their work. If the device works with a list and it is found on the disc, playback will proceed in the order specified in the list. Otherwise, playback will proceed in the order in which the tracks were recorded on the disc. If you want to use the playlist in the future, check the appropriate box. Once the desired tracks are marked and the saving options are set, click the button Save (Save). The process of saving tracks on the hard drive will begin, accompanied by the appearance of a dialog with a progress indicator. At the end of the process of saving tracks, a dialog will appear indicating that the saving was successful. Click the button OK

to close this dialog, and the program is ready for further work.

The dimples (striations) form a single helical track (in each layer) with a spacing of 0.74 microns between turns, which corresponds to a track density of 1,351 turns per millimeter. In total, this amounts to 49,324 turns, and the total length of the track reaches 11.8 km. The track is divided into sectors, each of which contains 2,048 bytes of data.

. Marks the end of a data area. The end zone sectors contain only 00h values. If a disc has two recording layers and is written in Opposite Track Path (OTP) mode, where the second layer starts from the outside of the disc and is read in the opposite direction to the first layer, this zone is called

Typically, a standard DVD's spiral track begins with the zero region and ends with the end (middle) zone, located 58.5 mm from the center of the disc or 1.5 mm from its outer edge. The length of one spiral path reaches 11.84 km. When reading the outer part of the track using a 20x CAV drive, which has a constant angular velocity (CAV), the data moves relative to the laser at a speed of 251 km/h. And despite such a high speed of data movement, the laser sensor accurately reads bit values ​​(valley/pad transitions), the dimensions of which do not exceed 0.4 microns.

There are single-layer and double-layer, as well as single-sided and double-sided versions of DVDs. Double-sided discs are essentially two single-sided discs glued back to back. There is a more significant difference between the two- and single-layer versions. The length of the grooves (stripes) of dual-layer discs is slightly longer, which leads to a slight decrease in the capacity of the disc.

Res. 6. DVD disc areas (sectional view)

The spiral track is divided into sectors, the repetition rate of which when reading or writing is 676 sectors per second. Each sector contains 2,048 bytes of data. The sectors are organized into data frames containing 2,064 bytes, of which 2,048 bytes are general data, 4 bytes contain identification information, 2 bytes are ID error detection (IED) codes, 6 bytes are media copyright information, and 4 bytes are represent the error detection code (EDC) for the data frame.

Data frames containing error correction code are converted into ECC frames. Each ECC frame contains a 2,064-byte data frame, as well as 182 bytes of parity parity (PO) and 120 bytes of parity parity (PI), for a total of 2,366 bytes for each ECC frame.

ECC frames are converted into separate 91-byte groups into physical disk sectors. This is done using the 8/16 modulation method, in which each byte is converted into a special 16-bit value selected from a table. These 16-bit values ​​are designed to contain no less than 2 or more than 10 contiguous bits that have a value of zero (0). This form of coding with a limitation on the length of the record field (Run Length Limited - RLL) is called the RLL 2.10 scheme. Once the conversion is complete, 320 bits (40 bytes) of synchronization data are added to each frame. Thus, after converting the ECC frame to a physical sector, the total number of bytes in the sector reaches 4,836.

Digital Versatile Discs, unlike standard CDs, do not use subcodes. Instead, each data frame contains identification bytes (ID) used to store the sector number and other sector-specific information.