Dec 102017
Seagate's low-level verify/format program works on some IDE drives.
File SGTFMT2.ZIP from The Programmer’s Corner in
Category HD Utilities
Seagate’s low-level verify/format program works on some IDE drives.
File Name File Size Zip Size Zip Type
README2.1ST 992 526 deflated
SGATFMT2.DOC 64022 21852 deflated
SGATFMT2.EXE 105517 47441 deflated
TPCREAD.ME 199 165 deflated

Download File SGTFMT2.ZIP Here

Contents of the SGATFMT2.DOC file

Ver. 2.0
By R. Dimick
R. Stacy
(tm) B. Rudock

(C)opyright 1992, Seagate Technology, Inc.
Scotts Valley, California USA
Tech Support BBS (408)438-8771

Introduction (see License agreement at the end of this document)

SGATFMT2 (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.)

SGATFMT2 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

SGATFMT2 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.)

SGATFMT2 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, SGATFMT2 will exit and issue an appropriate debug
command to initiate the controller's built in lo-level format.

SGATFMT2, in this v2.0 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


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.

SGATFMT2 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:

Ver. 2.0
By R. Dimick
R. Stacy
(tm) B. Rudock

> Drive 0

< (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

Drive 0
> ST124

< (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:


1. Format Drive
* 2. Enter Defects
3. Verify Drive
4. Format/Verify Drive
5. Choose Another Drive
* 6. Optimize Interleave

(* Menu option not available when AT/IDE ZBR drive selected.)

1. Format Drive : This is the "meat and potatoes" part of
SGATFMT2. 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 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 SGATFMT2.

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 SGATFMT2 without a BFI (a BFI of 0),
then the entire track is locked out (see ANATOMY OF A SECTOR below).
Once all of the defects are entered, the specific areas will be marked
as bad upon exiting the module.

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 provides,
however, for marking out "discovered" defects at the sector level 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. SGATFMT2 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.


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 switch time * SPT * spindle speed ) / 60,000 ] + 2

Ex: [( <15 S * 17 * 3600 ) / 60,000 ] + 2 = 2

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:


[( 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. For all intents
and purposes, 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 remind you 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.



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 SGATFMT2. 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. SGATFMT2 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 SGATFMT2 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, SGATFMT2 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., SGATFMT2 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

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!), SGATFMT2 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.



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 landmark 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.


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

Anatomy of a Sector (17-sector, 512 byte/sector):
Index Index

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Gap1 Gap4

571 Bytes Total

(Field Types:)
Sync.ID FieldGap2Data FieldGap3
1 2 3 4 5 6 7 8 9 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
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
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
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

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.


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 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.


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?

2. Does the CMOS have a drive type that has the same number of
formatted megabytes?

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.

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.

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

*** 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.



(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.

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.

The time difference between the leading edge of read and the
center of the data window.

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.

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.

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

Purely an operating system function or term describing the number
of sectors that the operating system allocates each time disc space is

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.

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.

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.

(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.

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.

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.

Information processed by a computer, stored in memory, or fed into
a computer.

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.

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.

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.

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.

Any of the circular tracks magnetized by the recording head during
data storage.

(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.

A complete disc surface is dedicated for servo data.

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.

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).

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.

(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.

Servo data is embedded or superimposed along with data on every

(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.

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.

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.

A disc drive with discs that cannot be removed from the drive by
the user, e.g., WINCHESTER DISC DRIVE.

Location on the data track, where the direction of magnetization
reverses in order to define a 1 or 0 bit.

(FCI). Linear recording density defined as the number of flux
changes per inch of data track.

Frequency modulation CODE scheme, superceded by MFM, which is
being superceded by RLL.

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.

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.

(flux changes per inch), also FRPI, the number of Flux Reversals
per inch.

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.

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.

The narrower the gap width, the closer the tracks can be placed.
Closer track placement results in higher TPI.

GROUP CODE ENCODING. Data encoding method.

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.

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).

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.

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

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.

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.

(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.

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.

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.

The ratio of physical disc sectors skipped for every sector
actually written.

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.

(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.

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.

A media defect can cause a considerable reduction of the read
signal (missing pulse or DROP-OUT), or create an extra pulse

One million bytes (exactly 1,000,000 bytes). Abbreviation: MB or

(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).

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.

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.

To access a storage location and obtain previously recorded data.

Return to Track Zero. A common disc drive function in which the
heads are returned to track 0 (outermost track).

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.

With regards to magnetic recording, the band width (or frequency
response) of the recording heads.

(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".

(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.

The speed at which the media spins. On a 5-1/4 or 3-1/2"
Winchester drive it is usually 3600 rpm.

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 allows any sector with a defect to be mapped and
bypassed. The next contiguous sector is given that sector address.

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.

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.

A bit error during playback which can be corrected by repeated
attempts to read.

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.

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

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.

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.

A certain part of each CYLINDER contains servo positioning data.
Gap spacing between each sector contains servo data to maintain
position on that cylinder.

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



Available on the Seagate Tech Support BBS (408)438-8771:

Specifications and jumper drawings for all Seagate Disc Drives and

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.



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

End of License agreement.


 December 10, 2017  Add comments

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>