шъъъшш шъъъшш шъъъшшъъъшъъъъ шъъъъъъ шъъшъъшшъъъшъъъъ Ver. 3.0 шэээээ шш шэээшш шш шээээ шш шш шш шш шш шш ъшш шш шш шш шш шш шш шш шэээшш шэээшш шш шш шш шш шш шш шш (tm) (C)opyright 1993, Seagate Technology, Inc. Scotts Valley, California USA Tech Support BBS (408)438-8771 Introduction (see License agreement at the end of this document) ---------------------------------------------------------------- ! READ THIS ENTIRE DOCUMENT BEFORE USING THIS PROGRAM. THIS PROGRAM ! IS DESTRUCTIVE TO USER DATA. SEVERAL SPECIFIC WARNINGS AND ! RECOMMENDATIONS ARE GIVEN THAT MAY PERTAIN TO YOUR DISC DRIVE. SGATFMT3 (Seagate Format) is a lo-level formatting utility designed for AT 286/386/486 systems, only. (If the program is run on an XT, most likely a stack overflow error message will display.) SGATFMT3 does not use the system BIOS to access the drive, but instead uses the AT register command set. This means that it is not necessary to pre-set a CMOS drive-type prior to the lo-level format. The CMOS drive type will become mandatory, however, prior to partitioning and the DOS hi-level format (see the section below on SETTING CMOS DRIVE TYPES). SGATFMT3 only works if the controller/host adapter is set to the primary hard drive port addresses of 1F0-1F7. (This is the common port address used on most controllers.) SGATFMT3 checks to see if a Seagate ST21/22 M or R controller is installed with its on-board controller bios enabled. If this condition exists, SGATFMT3 will exit and issue an appropriate debug command to initiate the controller's built in lo-level format. SGATFMT3, in this v3.0 release, is designed and LIMITED to work with the following Seagate disc drive interfaces: ST412 (both MFM and RLL), ESDI (with controller bios disabled), and AT/IDE (with certain limitations). SCSI interface disc drives are not supported. (See the section "ABOUT DRIVES NOT LISTED") ========================================================== There are three basic steps to preparing a hard disc drive for use in a computer system: 1. Lo-level format (MFM, RLL, ESDI, and some SCSI) 2. Partitioning with the operating system software. 3. Hi-level formatting with the operating system software. SGATFMT3 addresses step number 1. ========================================================== The opening first screen is used to determine which of two drives is to be selected for the lo-level format. If you only have one drive then select drive 0 by pressing 0, followed by the Enter key: шъъъшш шъъъшш шъъъшшъъъшъъъъ шъъъъъъ шъъшъъшшъъъшъъъъ Ver. 3.0 шэээээ шш шэээшш шш шээээ шш шш шш шш шш шш ъшш шш шш шш шш шш шш шш шэээшш шэээшш шш шш шш шш шш шш шш (tm) зддддддддддддддддддддддддддддддддд© Ё дддддддддддддддшддддддддддддддд Ё Ё дддддддддддддддшддддддддддддддд Ё Ё дддддддддддддддшддддддддддддддд Ё эээаээээээээээээээээшээээээээээээээээаэээ ш ш здддддддддддддддддддддддддддддд> Drive 0 ш Ё ш э ш Ё ш ш Ё ъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъ Ё юдддддддддддд< (Look for your choice to show up here) Please select physical hard drive 0 or 1 press to select After the drive selection is made, the next step is to identify the model: зддддддддддддддддддддддддддддддддд© Ё дддддддддддддддшддддддддддддддд Ё Ё дддддддддддддддшддддддддддддддд Ё Ё дддддддддддддддшддддддддддддддд Ё эээаээээээээээээээээшээээээээээээээээаэээ ш ш ш Drive 0 ш ш э зддддд> ST124 ш ш Ё ш ъъъъъъъъъЁъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъъ Ё юдд< (Look for your choice here) Please select a Seagate drive model, press to select U =prev D =next HOME =first END =last PGUP =U10 PGDN =D10 Once the model has been identified and the Enter key is pressed , the Main Menu appears: иммммм FORMAT OPTIONS ммммммм╩ ╨ ╨ ╨ 1. Format Drive ╨ * ╨ 2. Enter Defects ╨ ╨ 3. Verify Drive ╨ ╨ 4. Format/Verify Drive ╨ ╨ 5. Choose Another Drive ╨ * ╨ 6. Optimize Interleave ╨ ╨ ╨ хмммммммммммммммммммммммммммм╪ (* Menu option not available when AT/IDE ZBR drive selected.) 1. Format Drive : This is the "meat and potatoes" part of SGATFMT3. When this selection is made a warning appears, letting you know that ALL DATA WILL BE ERASED. This is very serious business! If you haven't backed up your data, then STOP! Under no circumstances is Seagate responsible for lost data. If you elect to go on, you will be asked to select or test for the proper interleave value. Next you will be queried for head and cylinder skew values (see INTERLEAVE and SKEWING sections below). A format on a disc drive is very controller dependent and usually means that the format performed by one controller cannot be utilized by another. 2. Enter Defects : Affixed to the top of every Seagate MFM and RLL disc drive, is a list of micro-defects that were found to exist at the time of manufacture. Seagate's original list should contain less than 1 defect per formatted megabyte and defect-free on the first two cylinders. The micro-defects that have been detected are generally of two types: hard and soft. A hard defect is usually a surface problem and a soft defect is usually a magnetic anomaly of some kind. Soft defects are discovered at the factory with very sophisticated test equipment, while hard defects can be discovered with conventional software like SGATFMT3. The typical defect label on the top of the drive is usually made up of three columns: Cyl Hd BFI and might look like this: 67 0 7814 68 0 7815 69 0 7816 175 2 3316 and so on. The column heading "BFI" stands for Bytes From Index. It may also be listed as "BCAI" which stands for Byte Count After Index, and is the same thing. The Index pulse is usually generated by a Hall sensor that is imbedded in the spindle motor or else it is encoded on servo tracks. This index pulse is considered the absolute point of reference for the BFI or BCAI count. With BFI, an individual sector can be located and locked out as opposed to locking out the entire track. If a defect is entered in SGATFMT3 without a BFI (a BFI of 0), then the entire track is locked out . Once all of the defects are entered, the specific areas will be marked as bad upon exiting the module. (see ANATOMY OF A SECTOR below) 3. Verify Drive : This module should still be proceeded by a complete backup before use. Verify is available to search out hard defects. If the micro-defect list has been removed from the drive or the suspicion of a new defect arises, then Verify can be run. It will report to the screen, and optionally to the printer, a cylinder, head, and sector reference. Unfortunately, a specific BFI cannot be reported. Therefore, if a subsequent lo-level format is performed, a BFI of 0 will need to be entered. Verify will ask if you want to do destructive pattern testing. If answered "No", the program operates in a read-only mode. If answered "Yes", you can choose up to nine different patterns that are used in write-read mode. (Note: a high capacity drive may take several hours to complete if all nine patterns are selected.) 4. Format/Verify Drive : This function combines the Format and Verify procedures into a single operation. This step does provide, however, for marking out "discovered" defects at the sector level instead of whole tracks at the time of formatting. 5. Choose Another Drive : If two physical drives are installed, this allows for switching between them. Be ABSOLUTELY SURE you are aware of which drive is selected. The next saddest person in the world is the one who formats the wrong drive! (Chin up.. worse things can happen.) 6. Optimize INTERLEAVE : The interleave value for a hard disc drive determines how many times a disc needs to spin in order to read a single track of data. The typical disc drive usually spins at 3,600 rpm (or 60 times per second). On a MFM disc drive with 17 sectors per track, the Read/Write heads, drive circuitry, controller and CPU are required to process all 17 sectors in 1/60th of a second. SGATFMT3 can test the system and report which interleave yields the fastest data transfer rate for your system (this is a data destructive test, be sure to back up 100% of your data before running the interleave tests). The best interleave possible is 1 to 1, meaning 1 revolution to read 1 track of data. Interleaves are always whole numbers, so the next best interleave is 2 to 1. 1 to 1: 1- 2- 3- 4- 5- 6- 7- 8- 9-10-11-12-13-14-15-16-17 (with sector 17 looping around to meet sector 1) 2 to 1: 1-10- 2-11- 3-12- 4-13- 5-14- 6-15- 7-16- 8-17- 9 (with sector 9 looping around to meet sector 1) It takes a little getting used to looking at this, but the most important fact to keep in mind is that the operating system reads the sectors in sequential order and will read on until the next sector in sequence appears. On the 2 to 1 interleave example the disc will need to spin two times in order to read all 17 sectors. Most of today's modern controllers are designed for a 1 to 1 interleave. Some early 16-bit controllers for 286's were only 3 to 1 or 2 to 1. An interesting problem happens if a 1 to 1 interleave is selected on a controller not designed for this speed: The Disc ends up performing like it has a 17 to 1 interleave! The reason for this is quite simple. If sector 2 immediately follows sector 1, and the controller isn't ready to read sector 2, then the disc needs to spin all the way around again in order to pick up on sector 2. This extra spin would be needed for all 17 of the sectors. ========================================================== SKEWING -------- By way of an analogy, the function of the modern disc drive has been described like this: "Today's new generation of disc drives achieve the engineering equivalent of a Boeing 747 flying at MACH 4 just two meters above the ground, counting each blade of grass as it flies over. The read/write head floats at 12 millionths of an inch above the surface of the disc which is turning at 3,600 revolutions per minute. Read/write heads position precisely over information tracks which are 800 millionths of an inch apart and the data is electronically recorded at 20,000 bits per inch." Skewing is best understood by first looking at the layout of a non-skewed disc drive. With the limitations of a two-dimensional drawing, a single circular MFM track has 17 sectors and would look like this: 1- 2- 3- 4- 5- 6- 7- 8- 9-10-11-12-13-14-15-16-17 (with sector 17 looping around to meet sector 1) The platters within the drive are spinning at a very high rate (usually 3,600 rpm), so one sector is passing beneath the R/W head once every 980 millionths of a second! This is obviously a very small timing window. When the entire track is processed, it is time to move to the next head (on another surface) in the cylinder. For example: a drive with two heads reads track 1 head 1, track 1 head 2, then repositions the heads over the next track and reads track2 head 1, track 2 head 2, and so on. The time it takes to switch between heads is extremely fast since it is an electronic change. The time it takes to reposition over another cylinder, however, takes significantly longer since it requires a mechanical movement that is an order of magnitude slower. Looking again at the 17 sectors, if we stack two heads we see: head 1 : 1- 2- 3- 4- 5- 6- 7- ...... -17 head 2 : 1- 2- 3- 4- 5- 6- 7- ...... -17 We would expect sector 1 on head 2 to immediately follow sector 17 on head 1. Unfortunately, this doesn't happen because it TAKES TIME (or "overhead") to switch to the new head, and by the time it does, sector 1 has already gone by! Therefore the R/W head waits for the disc to spin around once for sector 1 to show up again so it can get on with its job. Effectively, we have wasted one disc revolution that equals 1/60th of a second which could have processed almost an entire track of 17 sectors. This is the crux of the problem that skewing addresses: eliminating unnecessary disc revolutions. The solution is easy; shift the beginning position of sector 1 head 2 enough to compensate for the head switching overhead. That way when head 1 sector 17 finishes and the head switches, sector 1 head 2 would be spinning into place. Remembering that tracks are circular, it would look like this: head 1 : 1- 2- 3- 4- 5- 6- 7- ...... -15-16-17 head 2 : 16-17- 1- 2- 3- 4- 5- ...... -13-14-15 \--|--/ | Shifting these two sectors gives us time to allow for the head switching overhead and is the equivalent to HEAD SKEW = 2. In normal use, a disc drive switches heads many times more often than it does switching physical cylinders. The data throughput can rise dramatically when a head skew is in place. For example, a simple non-head skewed MFM drive might have a transfer rate of 380kps and the transfer rate of a drive with a head skew of 2 could rise to around 425kps. (Since we've listed a kind of performance result, here, it is VERY important to point out that ALL systems/controllers have different amounts of overhead and processing power, not to mention the wide range of results from different transfer rate diagnostics. See the section ABOUT TRANSFER RATES below.) A formula for calculating a head skew value is as follows (but be sure to read on): HEAD SKEW = [( head switch time * SPT * spindle speed ) / 60,000 ] + 2 Ex: [( <15 ФS * 17 * 3600 ) / 60,000 ] + 2 = 2 юддддддддбддддддддддддддддддддды а Basically, this evaluates to zero, and the 2 is a typical overhead for most MFM controllers. Cylinder skewing is usually a little more drastic. It stands to reason that since the mechanics of repositioning the head assembly is going to be significantly slower than an electronic head switch, the value for a cylinder skew will be larger. Going back to our two head drive, we might see: Cyl 1: head 1 : 1- 2- 3- 4- 5- 6- 7- 8- 9-10-11-12-13-14-15-16-17 head 2 : 16-17- 1- 2- 3- 4- 5- 6- 7- 8- 9-10-11-12-13-14-15 Cyl 2: head 1 : 8- 9-10-11-12-13-14-15-16-17- 1- 2- 3- 4- 5- 6- 7 \-----------|-----------/ | Shifting these eight sectors gives us time to account for the cylinder switching overhead and is the equivalent to CYLINDER SKEW = 8. A formula for calculating a cylinder skew value is as follows: CYLINDER SKEW = [( max track to track time * SPT * spindle speed ) / 60,000 ] + OHFactor Ex: [( 8 msec * 17 * 3600 ) / 60,000 ] + 0 = 8 (ok to round down on MFM) Note: OHFactor is an 'overhead factor' that is tied to SPT or sectors per track. After some casual experimentation, we've figured - SPT OHFactor 17 0 or 1 (usually MFM drives) 26 - 31 1 or 2 (usually RLL drives) 33 - 52 2 or 3 (usually ESDI drives) 53 - >> 3 or 4 (usually high end ESDI drives) The "0 or 1" type values are intended to be ambiguous, and are meant to illustrate that these values are system/controller dependant. The higher of the two numbers is the most conservative. Generally, choosing a value a little high is not as bad as choosing a value too low, thereby causing a wasted disc revolution. Now is a good time to recall that it is the head skew value that offers the most significant boost to the transfer rate, while an optimized cylinder skewing helps only when the heads are repositioned over a different track. If you use a transfer rate utility to measure performance results, be advised that many of them just use a single cylinder and don't reflect cylinder skews. ========================================================== ABOUT DRIVES NOT LISTED ----------------------- Some points about lo-level formatting drives not listed above: In the case of all SCSI drives: These drives use a controller (properly called a host adapter) that has an onboard BIOS chip. Coded within this bios chip is a lo-level format utility (called 'firmware' as opposed to 'software') which can initiate special SCSI commands. The fact that virtually all SCSI host adapters have this capability, precludes the need for a stand-alone software utility like SGATFMT3. Defect management on SCSI drives is handled at the factory and/or by the drive "on-the-fly" on more advanced drives, and is transparent to the user. Access to the SCSI host adapter's lo-level format utility is usually through the DOS'S DEBUG utility. Typically, you would start DEBUG, and then at the "hyphen prompt" (DEBUG's user-friendly interface), type "G=C800:5" without quotes and followed by ENTER (where C800 is the BIOS upper memory address selected by jumpers on the host adapter). In the case of ESDI drives: These drives normally use a controller with an onboard BIOS that has the lo-level utility. Many ESDI drives have cylinder counts that exceed the DOS limitation of 1024. The ESDI controller's on-board bios is required to "translate" these values in order to achieve full capacity from the drive. Defect management for ESDI drives has been simplified over that of typical MFM drives. The manufacturer has placed a small file on the drive which lists the coordinates of the defects (cylinder, head, and BFI or BCAI) that can be read by the controller, thereby eliminating the need to enter them by hand. Access to the ESDI controller's lo-level format utility is usually through the DOS'S DEBUG utility. Typically, you would start DEBUG, and then at the "hyphen prompt" (DEBUG's user-friendly interface), type a GO command, -G=C800:5 (where C800 is the BIOS upper memory address selected by jumpers on the controller). ESDI drives can be defined optionally, with the BIOS on the controller card disabled, in a user-definable or custom CMOS drivetype. SGATFMT3 supports this bios-disabled condition. In the case of RLL drives : These drives also normally use a controller like the ST21/22R controllers with an onboard BIOS that has the lo-level utility. Defect management for RLL drives is the same as MFM drives. Defects are usually listed on a sticker affixed to the top of the drive and need to be entered manually during the lo-level format. Access to the RLL controller's lo-level format utility is usually through the DOS'S DEBUG utility. Typically, you would start DEBUG, and then at the "hyphen prompt" (DEBUG's user-friendly interface), type a GO command, -G=C800:5 (where C800 is the BIOS upper memory address selected by jumpers on the controller). RLL drives can be defined optionally, with the BIOS on the controller card disabled, in a user-definable or custom CMOS drivetype. This version of SGATFMT3 supports RLL drives that are fully defined in CMOS with the controller BIOS disabled. In the case of AT (IDE) drives: AT (IDE) drives can be divided into three separate scenarios: Early, Swift and ZBR. 1. EARLY: When AT interface drives (aka IDE - integrated drive electronics, but so are SCSI's) were first introduced (ST157A family), we strongly warned and cautioned against any attempt to lo-level format the drives because 1) the factory written defect-mapping files might be erased on reserved areas of the drive, and 2) the optimized interleave and skewing values used would be forfeited giving slow transfer rates. At this stage of development, SGATFMT3 lists these drives only as a fall back option, in lieu of a factory repair format. If the drive has somehow lost its original format, or the partition structure been corrupted by a virus etc., SGATFMT3 could be used to reformat _without_ the benefit of the defect mapping files. Any defects will need to be "rediscovered" again; first, by the DOS high level format and second, by a third-party disk scanning utility. These utilities are quite likely to locate all of the hard errors, but unlikely to find the soft errors. The only way to completely evaluate a drive for both hard and soft error is by a factory repair with extremely sophisticated diagnostic equipment. (See the glossary section for HARD and SOFT ERRORS.) 2. SWIFT: As the AT interface products became more sophisticated with new technology and the introduction of the Swift drives (models like ST1239A, ST1201A etc), lo-level formatting became pretty much "half" of a problem. When these drives are in translation mode (non-physical geometry definitions), a lo-level format is harmless to the factory defect-mapping files and optimized skewing (albeit destructive to user data) since it doesn't re-sector the drive. If, however, the Swift drive is in true physical mode, then the lo-level format will re-sector the drive. 3. ZBR: Finally, today's AT interface drives (like the ST-1144A and ST-3144A) are often Zone Bit Recorded (ZBR). ZBR drives, have variable sectors per track, depending on the zone of the drive. The outside tracks, being larger in circumference (i.e. track length is longer), are able to hold more sectors than the innermost tracks. In this scenario, it is IMPOSSIBLE to define the drive in CMOS setup with true physical values. Cylinders and heads, yes.... but not the sectors per track. Therefore, these drive are ALWAYS in translation mode and immune to a re-sectoring lo-level format. On ZBR AT interface drives (Seagate, at least... others UNK), the factory defect mapping files are fully protected, and since the drive is always in translation, the optimized skewing is also protected. As to defect management, most AT interface drive's show 0 bytes in bad sectors under CHKDSK. This is a courtesy reallocation or "slipping" of bad sectors by the factory format, and not part of the interface definition. There are a few good reasons to consider a lo-level format for a ZBR AT/IDE drive. Because a lo-level will "data-scrub" all the sectors, this may be the only way to delete a corrupted partition record, or partition record from another operating system, or even a virus infection. If a new defect surfaces, maybe from a head slap (earthquake!), SGATFMT3 is able to find and lock out the offending sector, provided the defect is not in the ID portion of the sector. In this method, a kind of mid-level format, the locked out sector will be found again during the DOS hi-level format and will indicate as "bytes in bad sectors" at the conclusion. ========================================================== ABOUT TRANSFER RATES There seems to be a lot of confusion concerning data transfer rates on hard disk drives. This is a pity, as this should be a very straightforward issue. The first thing to do is forget the sales literature in expressing the practical transfer rate of a drive. The internal and external transfer ratings are only useful as an estimate of the maximum bus transfer rate of the area in question. What that usually means is that those rates are the measure of the speed both data and commands can be transferred across a given bus in a given rate of time. For all practical intents and purposes, this is only a valid for clocking command transfer rates, and data transfer in burst mode. For sustained data transfer rate, the bottom line is, the more sectors that pass under the head in a second, the faster the data comes off of the drive. To calculate the sustained rate, use this formula : (512 * Drive RPM * SPT) / (Interleave * 60) This rating is in Bytes / Second. For Example, a 251 at 3:1 interleave would transfer data as follow : (512 * 3600 * 17)/(3*60)=174,080 Bytes/second. This is the maximum data transfer rate possible without caching. To differentiate, and explain failings, you must realize that the above formula is for IDEAL conditions. Delays can be introduced by track crossings, head switch time, or, most importantly, how the system asks for the data. There is also the system overhead to look at, which can be grouped in with data inquiry delay. To illustrate the latter, think of the drive rotating at 3600 RPMs. The host system wants several sectors worth of information for its spreadsheet. It asks for a sector read. The drive acknowledges the command. the system waits. The drive steps to the proper track. The drive reads. The host acknowledges. The host asks for the next sector. The drive, which has been spinning all this time as drives do, no longer has its heads over that sector, because the host didn't ask for data in time. The drive spins. The sector is read, and so on. This procedure is much faster if the host just asks for a multiple sector read, as once the data is located, it streams directly off of the drive. This condition can be masked by the use of buffers, because the next few data requests can be satisfied by the queue, or buffer, whether built into the drive controller, or allocated to the system memory. Both of these schemes anticipate a multiple sector read beforehand, and fill memory locations with the data from the next few contiguous sectors. Although this works for the most part, once the queue is exhausted, we are back to the limitation of the sustained transfer rate, to be found by the aforementioned formula. ========================================================== ANATOMY OF A SECTOR ------------------- The purpose of a track format is to organize a data track into smaller sequentially numbered blocks called sectors. The beginning of each sector is defined by a pre-written identification (ID) field which contains the Logical sector address plus cylinder and head information. The ID field is then followed by a user supplied data field. Anatomy of a Sector (17-sector, 512 byte/sector): Index Index з© з© ЁЁ ЁЁ ыюдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддбдддыюд Ё 1 Ё 2 Ё 3 Ё 4 Ё 5 Ё 6 Ё 7 Ё 8 Ё 9 Ё10 Ё11 Ё12 Ё13 Ё14 Ё15 Ё16 Ё17 Ё Gap1 Ё Ё Gap4 Ё Ё 571 Bytes Total здддддддддддддддддддддддддддды юдддддддддддддддддддддддддддддддддддддд© (Field Types:) зSync.ддбID FieldддддддддбGap2дддбData FieldддддддддддддддддддддддбGap3д© Ё 1 Ё 2 3 4 5 6 7 Ё 8 9 Ё 10 11 12 13 14 15 (Field No.) Field No. Bytes Field Description 1 13 ID VFO Lock A field of all zeros to synchronize the VFO for the ID. 2 1 Sync. Byte A1h with a dropped clock to notify the controller that data follows. 3 1 Address Mark FEh: ID data field follows. 4 2 Cylinder Address A numerical value in Hex defining the detent position of the actuator. 5 1 *Head Number A numerical value in Hex defining the head selected. 6 1 Sector Number A numerical value in Hex defining the sector for this section of the rotation. 7 2 **CRC Cyclic Redundancy Check information used to verify the validity of the ID information field just read. 8 3 Write Turn On Zeros written during format to Gap isolate the write splice created. This field assures valid reading of field number seven and allows the 13 bytes required for data VFO lock. 9 13 Data Sync. A field of all zeros to sync the VFO Lock VFO for the data field. 10 1 Sync. Byte A1h with a dropped clock to notify the controller that data follows. 11 1 Address Mark F8h: User data follows. 12 512 Data User Data. 13 2 **CRC Cyclic Redundancy Check information used to verify the validity of the user data field just read. 14 3 Write Turn Off Zeros written during update to Gap isolate the write splice created. This field assures valid reading of field number 13 and allows the 13 bytes required for VFO lock for the ID field of the next sector. 15 15 Inter-Record Gap A field of 4Eh which acts as a buffer between sectors to allow for speed variation. Index : This is a signal which occurs once per revolution and it functions to indicate the physical beginning of the track. * Head Number : bits 0, 1, 2 = Head Number bits 3, 4 = '00' bits 5, 6 = Sector Size = '00' bit 7 = Bad Block Mark ** CRC : These codes are generated by the controller, and written on the media during formatting. Data integrity is maintained by the controller, recalculating and verifying the ID Field check codes when the ID Field is read. An acceptable polynomial is: 16 12 5 X +X +X +1 In the case of the Data Field CRC, instead of two bytes of Data CRC, the controller may implement a multiple byte Error Correction Code (ECC) Data Field integrity system. An ECC system provides the possibility of data field read correction as well as read error detection. The correction/detection ability is dependent on the code chosen and the controller implementation. Gap1 : Provides a head switching recovery period and controller decision making period, so when switching from one track to another, sequential sectors may be read without waiting the entire rotational latency time (additional time may be required on 1 to 1 controllers by adding a head skew). Gap2 : This gap follows the CRC bytes of the ID field and continues to the data field address mark. Written by the controller, it provides both a pad to ensure a proper recording and recovery of the last bits of the ID Field check codes and to allow time for controller decision making plus a byte for a write splice. The write splice will be created on the media as soon as the interface Write Gate is activated when performing a Data Field update function. Gap3 : Also known as the inter-record gap, this gap follows the CRC bytes of the Data area. In addition to similarities to Gap2, it also provides a means to accommodate variances in spindle speeds. A track may have been formatted while the disk is running slower than nominal, then write updated with the disk running faster than normal. Without a gap, or if the gap is too small, the sync bytes or ID field of the next sector could be overwritten. The actual size of this padding, initially provided by the format function, will vary, affected by on the disk rotational speed variations when the track was formatted and each time the Data Field is updated. Gap4 : This is the speed tolerance gap for the entire track. It is required to insure that the entire track can be formatted during an Index Pulse to Index Pulse Track Format operation. This Preindex gap will vary in actual size, depending on the disk rotational speed (+-0.5%) and write frequency tolerance (+-0.01%) at the time of formatting. ========================================================== About Choosing a Drive Type in an AT: The drive types for SCSI, RLL, and ESDI interface drives are generally easy to determine, especially the SCSI drives. SCSI Almost all SCSI drives use DRIVE TYPE 0 or NONE, as the host adapter bios and the drive communicate together to establish the drive geometry. The low-level formatting routines are accessed on the host adapter through DEBUG. After the low-level format, follow the instructions for your DOS version for partitioning and system format. Note: SCSI drives from the Seagate Wren and Swift families are already low-level formatted at the factory. RLL / ESDI RLL and ESDI drives are usually not represented at all in the internal drive tables and consequently the controllers for these drives have onboard a ROM BIOS which either contains its own internal list of choices for the interface or else provides the ability to dynamically configure (define) the controller to the specific geometry of the drive. In the case of the ESDI interface, the controller gets parameters directly from the drive with a mode sense equivalent command. Unlike the SCSI, the CMOS drive type should start at 0 or NONE at the start of the installation (low level format through DEBUG - consult your controller manual for instructions), but it may be reset to DRIVE TYPE 1 by the controller card. Many of the older AT's only provided 14 (MFM only) or so drive types to choose from in the CMOS. The middle-aged AT's usually have up to 46 (still usually only MFM) types. Some newer AT's have drive types which begin to include direct support for the popular RLL and ESDI drives. If you have this newer kind of CMOS then by all means pick the one that matches the drive and DISABLE the controller Bios. (Note: This may also disable the controller's caching feature). Likewise, most new machines have a "User Definable" or "Custom" drive type that can be created and saved in the CMOS, thus providing a standard drive type. "User Definable" drive types will usually not work with most non-MS/PC-DOS applications. A special note on ESDI and other drives that have more than 1024 cylinders. Since DOS cannot access cylinders above this 1024 limit, a translation scheme may be elected in the controller's bios. As the number of Logical Block Address (LBAs) is defined as CYLINDERS*HEADS*SECTORS PER TRACK, translations that equal the same number of LBAs with the cylinder count below the 1024 limit will be devised. The controller bios will need to be ENABLED in order to utilize translations schemes. (e.g. Many popular controllers increase the number of sectors and/or heads and decrease the # of cylinders to achieve an equivalent number of LBAs. See your controller manual for details.) After low-level formatting, follow the instructions for your DOS version for partitioning and system format. AT / IDE This idea of translation schemes bring us to the AT or IDE (Imbedded Drive Electronics) interface. These drives are intelligent in that they can use the geometry that represents their true physical parameters or else they can "mimic" other drive geometries (or translations) that equal or are very close to, but NOT exceeding, the same number of logical blocks. (Translated LBA's <= Native LBA's.) Many AT/IDE drives have physical cylinder counts that are greater than 1024. Therefore, for DOS users, it is necessary to utilize the translate feature by using a geometry that keeps the cylinder count below 1024. In order of preference, choose the first that fits your system: 1. Does the CMOS have a drive type that matches your drive? no? 2. Does the CMOS have a drive type that has the same number of formatted megabytes? no? 3. Does the CMOS have a "custom" or "user definable" drive type option you can use? If so, use a translation geometry to keep the cylinder count below the DOS 1024 limit. no? 4. Do you have the Disk Manager program to provide a software driven solution? The Disk Manager will run automatically to perform the partitioning and system format. no? 5. Pick the drive type that comes closest to, but not exceeding, the formatted capacity of your drive. The final formatted capacity of the drive will be equal to the drive type chosen. *** Warning! ALL AT drives from Seagate are already low-level formatted at the factory. MFM (ST412 interface) Finally, the MFM drives and their associated drive types are next. If the internal drive type table lists the exact geometry, great. If not, then check to see if a "Custom" or "User Definable" CMOS option is available. Also, some AT 16-bit MFM controllers provide an onboard BIOS which will allow the unique geometry of the drive to be dynamically configured (our Seagate ST21M/22M MFM controllers have this VALUABLE feature). Otherwise, a drive type match that is close but not exceeding either the cylinder or head values is the only choice left. An exact match in the head count is definitely preferred when getting a "close" match. When there is no direct match in the internal drive type tables, a partitioning program may be needed to provide a software driven translation solution in order to achieve full capacity. Keep in mind that the drive will only format out to the capacity of the chosen drive type when not using partitioning software. In the event that the ST412 Interface drive has more than 1024 cylinders, a partitioning program will be needed in order to achieve full capacity. ========================================================== GLOSSARY OF DISC DRIVE TERMINOLOGY ADDRESS (physical) A specific location in memory where a unit record, or sector, of data is stored. To return to the same area on the disc, each area is given a unique address consisting of three components: cylinder, sector, and head. CYLINDER ADDRESSING is accomplished by assigning numbers to the disc's surface concentric circles (cylinders). The cylinder number specifies the radial address component of the data area. SECTOR ADDRESSING is accomplished by numbering the data records (sectors) from an index that defines the reference angular position of the discs. Index records are then counted by reading their ADDRESS MARKS. Finally, HEAD ADDRESSING is accomplished by vertically numbering the disc surfaces, usually starting with the bottom-most disc data surface. For example, the controller might send the binary equivalent of the decimal number 610150 to instruct the drive to access data at cylinder 610, sector 15, and head 0. BIT DENSITY Expressed as "BPI" (for bits per inch), bit density defines how many bits can be written onto one inch of a track on a disc surface. It is usually specified for "worst case", which is the inner track. Data is the densest in the inner tracks where track circumferences are the smallest. BIT JITTER The time difference between the leading edge of read and the center of the data window. BIT SHIFT A data recording effect, which results when adjacent 1's written on magnetic discs repel each other. The "worst case" is at the inner cylinder where bits are closest together. BIT SHIFT is also called pulse crowding. BLOCK A group of BYTES handled, stored and accessed as a logical data unit, such as an individual file record. Typically, one block of data is stored as one physical sector of data on a disc drive. CLOSED LOOP A control system consisting of one or more feedback control loops in which functions of the controlled signals are combined with functions of the command to maintain prescribed relationships between the commands and the controlled signals. This control technique allows the head actuator system to detect and correct off-track errors. The actual head position is monitored and compared to the ideal track position, by reference information either recorded on a dedicated servo surface, or embedded in the inter-sector gaps. A position error is used to produce a correction signal (FEEDBACK) to the actuator to correct the error. See TRACK FOLLOWING SERVO. CLUSTER SIZE Purely an operating system function or term describing the number of sectors that the operating system allocates each time disc space is needed. CODE A set of unambiguous rules specifying the way which digital data is represented physically, as magnetized bits, on a disc drive. One of the objectives of coding is to add timing data for use in data reading. See DATA SEPARATOR, MFM and RLL. COERCIVITY A measurement in units of orsteads of the amount of magnetic energy to switch or "coerce" the flux change (di-pole) in the magnetic recording media. CONTROLLER A controller is a printed circuit board required to interpret data access commands from host computer (via a BUS), and send track seeking, read/write, and other control signals to a disc drive. The computer is free to perform other tasks until the controller signals DATA READY for transfer via the CPU BUS. CYCLIC-REDUNDANCY-CHECK (CRC). Used to verify data block integrity. In a typical scheme, 2 CRC bytes are added to each user data block. The 2 bytes are computed from the user data, by digital logical chips. The mathematical model is polynomials with binary coefficients. When reading back data, the CRC bytes are read and compared to new CRC bytes computed from the read back block to detect a read error. The read back error check process is mathematically equivalent to dividing the read block, including its CRC, by a binomial polynomial. If the division remainder is zero, the data is error free. CYLINDER The cylindrical surface formed by identical track numbers on vertically stacked discs. At any location of the head positioning arm, all tracks under all heads are the cylinder. Cylinder number is one of the three address components required to find a specific ADDRESS, the other two being head number and sector number. DAISY CHAIN A way of connecting multiple drives to one controller. The controller drive select signal is routed serially through the drives, and is intercepted by the drive whose number matches. The disc drives have switches or jumpers on them which allow the user to select the drive number desired. DATA Information processed by a computer, stored in memory, or fed into a computer. DATA ACCESS When the controller has specified all three components of the sector address to the drive, the ID field of the sector brought under the head by the drive is read and compared with the address of the target sector. A match enables access to the data field of the sector. DATA ADDRESS To return to the same area on the disc, each area is given a unique address consisting of the three components: cylinder, head and sector. HORIZONTAL: accomplished by assigning numbers to the concentric circles (cylinders) mapped out by the heads as the positioning arm is stepped radially across the surface, starting with 0 for the outermost circle. By specifying the cylinder number the controller specifies a horizontal or radial address component of the data area. ROTATIONAL: once a head and cylinder have been addressed, the desired sector around the selected track of the selected surface is found by counting address marks from the index pulse of the track. Remember that each track starts with an index pulse and each sector starts with an address mark. VERTICAL: assume a disc pack with six surfaces, each with its own read/write head, vertical addressing is accomplished by assigning the numbers 00 through XX to the heads, in consecutive order. By specifying the head number, the controller specifies the vertical address component of the data area. DATA FIELD The portion of a sector used to store the user's DIGITAL data. Other fields in each sector include ID, SYNC and CRC which are used to locate the correct data field. DATA SEPARATOR Controller circuitry takes the CODED playback pulses and uses the timing information added by the CODE during the write process to reconstruct the original user data record. See NRZ, MFM, and RLL. DATA TRACK Any of the circular tracks magnetized by the recording head during data storage. DATA TRANSFER RATE (DTR). Speed at which bits are sent: In a disc storage system, the communication is between CPU and controller, plus controller and the disc drive. Typical units are bits per second (BPS), or bytes per second, e.g., ST506/412 INTERFACE allows 5 Mbits/sec. transfer rate. DEDICATED SERVO SYSTEM A complete disc surface is dedicated for servo data. DISC/PLATTER For rigid discs, a flat, circular aluminum disc substrate, coated on both sides with a magnetic substance (iron oxide or thin film metal media) for non-VOLATILE data storage. The substrate may consist of metal, plastic, or even glass. Surfaces of discs are usually lubricated to minimize wear during drive start-up or power down. DROP-IN/DROP-OUT Types of disc media defects usually caused by a pin-hole in the disc coating. If the coating is interrupted, the magnetic flux between medium and head is zero. A large interruption will induce two extraneous pulses, one at the beginning and one at the end of the pin-hole (2 DROP-INs). A small coating interruption will result in no playback from a recorded bit (a DROP-OUT). ECC ERROR CORRECTION CODE: The ECC hardware in the controller used to interface the drive to the system can typically correct a single burst error of 11 bits or less. This maximum error burst correction length is function of the controller. With some controllers the user is allowed to the select this length. The most common selection is 11. ELECTRO-STATIC DISCHARGE (ESD) An integrated circuit (CHIP) failure mechanism. Since the circuitry of CHIPs are microscopic in size, they can be damaged or destroyed by small static discharges. People handling electronic equipment should always ground themselves before touching the equipment. Electronic equipment should always be handled by the chassis or frame. Components, printed circuit board edge connectors should never be touched. EMBEDDED SERVO SYSTEM Servo data is embedded or superimposed along with data on every cylinder. FCI (FLUX CHANGES PER INCH): Synonymous with FRPI (flux reversals per inch). In MFM recording 1 FCI equals 1 BPI (bit per inch). In RLL encoding schemes, 1 FCI generally equals 1.5 BPI. FILE ALLOCATION TABLE FAT: What the operating systems uses to keep track of which clusters are allocated to which files and which are available for use. FAT is usually stored on Track-0. FIRMWARE A computer program written into a storage medium which cannot be accidentally erased, e.g., ROM. It can also refer to devices containing such programs. FIXED DISC A disc drive with discs that cannot be removed from the drive by the user, e.g., WINCHESTER DISC DRIVE. FLUX CHANGE Location on the data track, where the direction of magnetization reverses in order to define a 1 or 0 bit. FLUX CHANGES PER INCH (FCI). Linear recording density defined as the number of flux changes per inch of data track. FM Frequency modulation CODE scheme, superceded by MFM, which is being superceded by RLL. FORMAT The purpose of a format is to record "header" data that organize the tracks into sequential sectors on the disc surfaces. This information is never altered during normal read/write operations. Header information identifies the sector number and also contains the head and cylinder ADDRESS in order to detect an ADDRESS ACCESS error. FORMATTED CAPACITY Actual capacity available to store user data. The formatted capacity is the gross capacity, less the capacity taken up by the overhead data used in formatting the discs. While the unformatted size may be 24 M bytes, only 20 M bytes of storage may actually be available to the user after formatting. FPI (flux changes per inch), also FRPI, the number of Flux Reversals per inch. GAP 1. FORMAT: Part of the disc format. Allows mechanical compensations (e.g. spindle motor rotational speed variations) without the last sector on a track overwriting the first sector. 2. HEAD: An interruption in the permeable head material, usually a glass bonding material with high permeability, allowing the flux fields to exit the head structure to write / read data bits in the form of flux changes on the recording media. GAP LENGTH Narrowing the head gap length achieves higher bit density because the lines of force magnetize a smaller area where writing data in the form of flux changes on the recording media. GAP WIDTH The narrower the gap width, the closer the tracks can be placed. Closer track placement results in higher TPI. GCR GROUP CODE ENCODING. Data encoding method. GUARD BAND 1. Non-recorded band between adjacent data tracks, 2. For closed loop servo drives, extra servo tracks outside the data band preventing the Carriage Assembly from running into the crash stop. HARD ERROR An error that occurs repeatedly at the same location on a disc surface. Hard errors are caused by imperfections in the disc surface, called media defects. When formatting hard disc drives, hard error locations, if known, should be spared out so that data ia not written to these locations. Most drives come with a hard error map listing the locations of any hard errors by head, cylinder and BFI (bytes from index - or how many bytes from the beginning of the cylinder). HARD ERROR MAP Also called defect map, bad spot map, media map. Media defects are avoided by deleting the defective sectors from system use, or assigning an alternative track (accomplished during format operation). The defects are found during formatting, and their locations are stored on a special DOS file on the disc, usually on cylinder 0. HEAD An electromagnetic device that can write (record), read (playback), or erase data on magnetic media. There are three types: Head Type BPI TPI Areal density Monolithic 8000 450 3.6 X 10 to 6th Composition 12000 1000 12 X 10 to 6th Thin-film 25000 1500 37.5 X 10 to 6th HEAD SLAP Similar to a head crash but occurs while the drive is turned off. It usually occurs during mishandling or shipping. Head slap can cause permanent damage to a hard disc drive. See HEAD CRASH. ID FIELD The address portion of a sector. The ID field is written during the Format operation. It includes the cylinder, head, and sector number of the current sector. This address information is compared by the disc controller with the desired head, cylinder, and sector number before a read or write operation is allowed. INDEX (PULSE): The Index Pulse is the starting point for each disc track. The index pulse provides initial synchronization for sector addressing on each individual track. INDEX TIME The time interval between similar edges of the index pulse, which measures the time for the disc to make one revolution. This information is used by a disc drive to verify correct rotational speed of the media. INTERFACE The protocol data transmitters, data receivers, logic and wiring that link one piece of computer equipment to another, such as a disc drive to a controller or a controller to a system bus. Protocol means a set of rules for operating the physical interface, e.g., don't read or write before SEEK COMPLETE is true. INTERLEAVE FACTOR The ratio of physical disc sectors skipped for every sector actually written. INTERLEAVING The interleave value tells the controller where the next logical sector is located in relation to the current sector. For example, an interleave value of one (1) specifies that the next logical sector is physically the next sector on the track. Interleave of two (2) specifies every other physical sector, three (3) every third sector and so on. Interleaving is used to improve the system throughout based on overhead time of the host software, the disc drive and the controller; e.g., if an APPLICATION PROGRAM is processing sequential logical records of a DISC FILE in a CPU time of more than one second but less than two, then an interleave factor of 3 will prevent wasting an entire disc revolution between ACCESSES. LATENCY (ROTATIONAL) The time for the disc to rotate the accessed sector under the head for read or write. On the average, latency is the time for half of a disc revolution. LOW LEVEL FORMAT The first step in preparing a drive to store information after physical installation is complete. The process sets up the "handshake" between the drive and the controller. In an XT system, the low level format is usually done using DOS's debug utility. In an AT system, AT advanced diagnostics is typically used. Other third party software may also be used to do low level format on both XTs and ATs. MEDIA DEFECT A media defect can cause a considerable reduction of the read signal (missing pulse or DROP-OUT), or create an extra pulse (DROP-IN). See HARD ERROR MAP. MEGABYTE One million bytes (exactly 1,000,000 bytes). Abbreviation: MB or Mbyte. MODIFIED FREQUENCY MODULATION (MFM). A method of recording digital data, using a particular CODE to get the flux reversal times from the data pattern. MFM recording is self-clocking because the CODE guarantees timing information for the playback process. The controller is thus able to synchronize directly from the data. This method has a maximum of one bit of data with each flux reversal. (See NRZ, RLL). NRZ NON-RETURN TO ZERO 1) User digital data bits; 2) A method of magnetic recording of digital data in which a flux reversal denotes a one bit, and no flux reversal a zero bit, NRZ recording requires an accompanying synchronization clock to define each cell time unlike MFM or RLL recording). No Seagate drives use NRZ recording methods. PRECOMPENSATION Applied to write data by the controller in order to partially alleviate bit shift which causes adjacent 1's written on magnetic media physically to move apart. When adjacent 1's are sensed by the controller, precompensation is used to write them closer together on the disc, thus fighting the repelling effect caused by the recording. Precompensation is only required on some oxide media drives. READ To access a storage location and obtain previously recorded data. RECALIBRATE Return to Track Zero. A common disc drive function in which the heads are returned to track 0 (outermost track). REDUCED WRITE CURRENT A signal input (to some older drives) which decreases the amplitude of the write current at the actual drive head. Normally this signal is specified to be used during inner track write operations to lessen the effect of adjacent bit "crowding." Most drives today provide this internally and do not require controller intervention. RESOLUTION With regards to magnetic recording, the band width (or frequency response) of the recording heads. RLL (RUN LENGTH LIMITED CODE). 1) A method of recording digital data, whereby the combinations of flux reversals are coded/decoded to allow greater than one (1) bit of information per flux reversal. This compaction of information increases data capacity by approximately 50 percent; 2) a scheme of encoding designed to operate with the ST412 interface at a dial transfer rate of 7.5 megabit/sec. The technical name of the specific RLL CODE used is "two, seven". ROM (READ ONLY MEMORY) A chip that can be programmed once with bits of information. This chip retains this information even if the power is turned off. When this information is programmed into the ROM, it is called burning the ROM. ROTATIONAL SPEED The speed at which the media spins. On a 5-1/4 or 3-1/2" Winchester drive it is usually 3600 rpm. SECTOR A sector is a section of a track whose size is determined by formatting. When used as an address component, sector and location refer to the sequence number of the sector around the track. Typically, one sector stores one user record of data. Drives typically are formatted from 17 to 26 sectors per track. Determining how many sectors per track to use depends on the system type, the controller capabilities and the drive encoding method and interface. SECTOR-SLIP Sector-slip allows any sector with a defect to be mapped and bypassed. The next contiguous sector is given that sector address. SERVO TRACK A prerecorded reference track on the dedicated servo surface of a closed-loop disc drive. All data track positions are compared to their corresponding servo track to determine "off-track/on-track" position. SKEWING Some low-level formatting routines may ask for a Head and/or Cylinder Skew value. The value will represent the number of sectors being skewed to compensate for head switching time of the drive and/or track-to-track seek time allowing continuous read/write operation without losing disk revolutions. SOFT ERROR A bit error during playback which can be corrected by repeated attempts to read. TRACK The radial position of the heads over the disc surface. A track is the circular ring traced over the disc surface by a head as the disc rotates under the heads. TRACK FOLLOWING SERVO A closed-loop positioner control system that continuously corrects the position of the disc drive's heads by utilizing a reference track and a feedback loop in the head positioning system. See also CLOSED LOOP. TRACK ZERO Track zero is the outermost data track on a disc drive. In the ST 506 INTERFACE, the interface signal denotes that the heads are positioned at the outermost cylinder. VOICE COIL MOTOR An electro-magnetic positioning motor in the rigid disk drive similar to that used in audio speakers. A wire coil is placed in a stationary magnetic field. When current is passed through the coil, the resultant flux causes the coil to move. In a disc drive, the CARRIAGE ASSEMBLY is attached to the voice coil motor. Either a straight line (linear) or circular (rotary) design may be employed to position the heads on the disc's surface. WEDGE SERVO SYSTEM A certain part of each CYLINDER contains servo positioning data. Gap spacing between each sector contains servo data to maintain position on that cylinder. WRITE CURRENT The optimum HEAD write current necessary to saturate the magnetic media in a cell location. ZBR (Zone Bit Recording) Trademark of Seagate Technology. A media optimization technique where the number of sectors per track is dependent upon the cylinder circumference. E.G. tracks on the outside cylinders have more sectors per track than the inside cylinders. The ZBR format is only done at the factory. These drives should not be low-level formatted by the end-user. ========================================================== Other ----- Available on the Seagate Tech Support BBS (408)438-8771: Specifications and jumper drawings for all Seagate Disc Drives and Controllers. Reprints of Installation Guides. Troubleshooting essays. FINDTYPE - Utility which displays bios drive type table and matches a Seagate model to the best drive type. Also prints complete specifications lists and much more! FINDINIT - Utility for Seagate controllers and host adapters that have onboard bios, namely ST01, ST02, ST05X, ST11M, ST11R, ST21M, ST21R, ST22M, and ST22R. Queries the system to determine bios memory address and initiates controller bios lo-level format. FLASHLED - TSR utility which shows disc drive activity on one of the keyboard LED's. DESK REFERENCE - Hypertext data system for all Seagate products, troubleshooting, other OEM phone numbers and much, much more. A must for dealers who do a fair amount of support for Seagate products. ========================================================== LICENSE AGREEMENT Seagate provides the accompanying object code software ("Software") and nonexclusively licenses its use on the following terms and conditions. The Software is copyrighted by Seagate. YOU ASSUME FULL RESPONSIBILITY FOR THE SELECTION OF THE SOFTWARE TO ACHIEVE YOUR INTENDED PURPOSES, FOR THE PROPER INSTALLATION AND USE. SEAGATE DOES NOT WARRANT THAT THE SOFTWARE WILL MEET YOUR REQUIREMENTS, THAT THE SOFTWARE IS FIT FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF THE SOFTWARE WILL BE ERROR FREE. SEAGATE EXPRESSLY DISCLAIMS ALL WARRANTIES, WHETHER ORAL OR WRITTEN, EXPRESSED OR IMPLIED, INCLUDING WITHOUT LIMITATION WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT WILL SEAGATE BE LIABLE TO YOU, YOUR CUSTOMERS OR OTHER USERS FOR ANY INDIRECT, INCIDENTAL, CONSEQUENTIAL, SPECIAL OR EXEMPLARY DAMAGES ARISING OUT OF OR IN CONNECTION WITH THE USE OR INABILITY TO USE THE SOFTWARE. End of License agreement. -=EOF: SGATFMT3.DOC=-