Restore A Quick Formatted Drive In Windows? Also Cost Of IDE RAID

On Tue, 31 Mar 2009 06:24:13 +1100, Franc Zabkar
put finger to keyboard and composed:

There are some claims on Steve Gibson's web site which I don't
understand.

Steve Gibson claims to be able to influence a HD's read AGC (automatic
gain control) circuit.

SpinRite's Defect Detection Magnetodynamics:
http://www.grc.com/srphysics.htm

He says that ..

"SpinRite is actually able to lower the amplification of the drive's
internal read-amplifier, then to cause the drive to encounter a
'minimum amplitude' data signal. If that data signal happened to fall
upon a spot of the drive's recording surface that is in any way 'weak'
then the drive will mis-read the data bit and return an error for the
sector containing the bad spot."

However US patent #6839193, entitled "Method and apparatus for
determining read-to-write head offset of a disk drive", assigned to
Seagate, shows that the data field of each sector is preceded by an
"AGC data field which may be used to set read sensitivity".

The patent states that "when the read/write head is over the AGC data
field the gain of the recovered signal is adjusted to provide a
predetermined constant amplitude signal in order to reduce
demodulation errors."

So it appears to me that the read amp's AGC circuit is "trained"
during the "preamble" to the data field and is therefore not
influenced by the data pattern. Of course I'm assuming that the
technology described in the above patent (dated Jan 4, 2005) is
actually being implemented in current products.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Just because ATA-4 retired it doesn't mean it is no longer implemented.
The original 4-byte ECC is long gone, but I recall an IBM manual
describing another bit (when set) will read the full ECC (400 bits or so).

"Franc Zabkar" wrote in message
...

I referred to my copy of IBM's Personal Computer AT Technical
Reference Manual. The Fixed Disk Adapter section talks about the Read
Sector command. In those days the Command Register's bit definitions
for this command were as follows:

7 6 5 4 3 2 1 0
---------------
0 0 1 0 0 0 L T

L = 0 for data only, 1 for data plus 4-byte ECC
T = 0 for retries enabled, 1 for retries disabled

This IDE reference talks about the above Read Long commands (22h and
23h):
http://www.repairfaq.org/filipg/LINK/F_IDE-tech.html

However, I don't see a corresponding command in today's ATA
specification. It was present in ATA-3 but seems to have been retired
in ATA-4.

http://www.t10.org/t13/project/d2008r7b-ATA-3.pdf
http://www.t10.org/t13/project/d1153r18-ATA-ATAPI-4.pdf

So you are correct -- today's drives do not appear to be able to
retrieve the ECC data. My apologies.

On Mon, 30 Mar 2009 15:28:17 -0700, "Eric Gisin"
put finger to keyboard and composed:

Just because ATA-4 retired it doesn't mean it is no longer implemented.
The original 4-byte ECC is long gone, but I recall an IBM manual
describing another bit (when set) will read the full ECC (400 bits or so).

So that must be what Steve Gibson means by "unique 'hardware level
access' (which no other utility has)".

Thanks for the correction.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Franc Zabkar wrote:
On Tue, 31 Mar 2009 06:24:13 +1100, Franc Zabkar
put finger to keyboard and composed:

There are some claims on Steve Gibson's web site which I don't
understand.

Steve Gibson claims to be able to influence a HD's read AGC (automatic
gain control) circuit.

SpinRite's Defect Detection Magnetodynamics:
http://www.grc.com/srphysics.htm

He says that ..

"SpinRite is actually able to lower the amplification of the drive's
internal read-amplifier, then to cause the drive to encounter a
'minimum amplitude' data signal. If that data signal happened to fall
upon a spot of the drive's recording surface that is in any way 'weak'
then the drive will mis-read the data bit and return an error for the
sector containing the bad spot."

If I remember correctly, that claim was bogus to begin with.
What he could do with MFM/RLL is to test with a pattern, that
was more sensitve to surface problems, as it used the weak spots
of the modulation. Pretty much like black magic, so he possibly
dumbed it down. The actuall read amplifier gain is not accessible
and was not accessible in the MFM/RLL times as well.

However US patent #6839193, entitled "Method and apparatus for
determining read-to-write head offset of a disk drive", assigned to
Seagate, shows that the data field of each sector is preceded by an
"AGC data field which may be used to set read sensitivity".

The patent states that "when the read/write head is over the AGC data
field the gain of the recovered signal is adjusted to provide a
predetermined constant amplitude signal in order to reduce
demodulation errors."

So it appears to me that the read amp's AGC circuit is "trained"
during the "preamble" to the data field and is therefore not
influenced by the data pattern. Of course I'm assuming that the
technology described in the above patent (dated Jan 4, 2005) is
actually being implemented in current products.

No idea.

Arno

On 31 Mar 2009 13:12:10 GMT, Arno put finger to
keyboard and composed:

Franc Zabkar wrote:
On Tue, 31 Mar 2009 06:24:13 +1100, Franc Zabkar
put finger to keyboard and composed:

There are some claims on Steve Gibson's web site which I don't
understand.

Steve Gibson claims to be able to influence a HD's read AGC (automatic
gain control) circuit.

SpinRite's Defect Detection Magnetodynamics:
http://www.grc.com/srphysics.htm

He says that ..

"SpinRite is actually able to lower the amplification of the drive's
internal read-amplifier, then to cause the drive to encounter a
'minimum amplitude' data signal. If that data signal happened to fall
upon a spot of the drive's recording surface that is in any way 'weak'
then the drive will mis-read the data bit and return an error for the
sector containing the bad spot."

If I remember correctly, that claim was bogus to begin with.
What he could do with MFM/RLL is to test with a pattern, that
was more sensitve to surface problems, as it used the weak spots
of the modulation. Pretty much like black magic, so he possibly
dumbed it down. The actuall read amplifier gain is not accessible
and was not accessible in the MFM/RLL times as well.

AIUI, Steve Gibson does not claim to be able to control the read amp
gain directly. AGC relies on a rolling average, so Gibson "trains" the
read amp with strong signals, causing the AGC to reduce the gain, and
then he feeds the amp with weak signals while the gain is still low.

As for whether AGC was used in the early MFM days, I have a circuit
diagram for the 20MB drive that was used in the original IBM PC/AT.
The read amp appears to consist of two-stage differential op-amps with
fixed gains, ie no AGC. The print is hard to read, but the op-amps
look like NE592 video amps which are also specified for use as floppy
disc head amplifiers.

Here is a datasheet:
http://www.datasheetarchive.com/pdf-...DSA-463326.pdf

However US patent #6839193, entitled "Method and apparatus for
determining read-to-write head offset of a disk drive", assigned to
Seagate, shows that the data field of each sector is preceded by an
"AGC data field which may be used to set read sensitivity".

The patent states that "when the read/write head is over the AGC data
field the gain of the recovered signal is adjusted to provide a
predetermined constant amplitude signal in order to reduce
demodulation errors."

So it appears to me that the read amp's AGC circuit is "trained"
during the "preamble" to the data field and is therefore not
influenced by the data pattern. Of course I'm assuming that the
technology described in the above patent (dated Jan 4, 2005) is
actually being implemented in current products.

No idea.

Arno

AIUI, the above patent talks about an AGC data field that is used to
train the servo amp so that it can accurately position the MR read
head. If a write is commanded, then the servo is jogged slightly
off-track to where the write coil is known to be. This may partly
explain why the seek time for writes is slightly greater than for
reads. Accurate positioning is dependent on accurate calibration of
the servo signal. For example, if an on-track read signal, after AGC,
is 10mV, then the write head's offset from the read head could be
where this signal drops to 8mV, say. AIUI these read-write head
offsets are calibrated at the factory and vary from track to track.

In addition to the above AGC data field, it appears that the user data
field has its own preamble of "training" bits which are laid down when
the data are written. In this case these training and sync bits are
used to condition the read amp when reading back the user data.

According to US patent number 6055118, entitled "holding automatic
gain control levels during read error recovery in a disc drive",
assigned to Seagate, the data block begins with a phase locked
oscillator (PLO) field which provides timing and amplitude information
for AGC and clocking. This is following by a training field and sync
field, the latter being used to identify the beginning of the user
data. The patent states that "the AGC first operates in an
'acquisition' mode during which large adjustments in the gain are made
in response to the initial signal levels provided by the preamp to
provide a coarse adjustment of the gain. The AGC then transitions into
a 'fine adjustment' mode, during which the AGC normally continues to
provide small, incremental adjustments to the gain as long as the read
channel is enabled as a particular sector is read".

So at first it appears that SpinRite may be able to influence the AGC
as claimed by Gibson. But then the patent goes on to describe what
happens when a read error is encountered. In this case the drive
re-reads the previous sector on the same track, and determines its
appropriate gain based on its PLO field. It then retries the bad
sector by turning off AGC and applying the *constant* gain derived
from the previous good sector. So it seems that any error that is a
consequence of an attempt to confuse AGC will be automatically
corrected by the drive on its second try, by *disabling* AGC.

BTW, the second patent applies to drives using PRML signal processing.

See http://www.google.com/patents?as_pnum=6055118

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Franc Zabkar wrote:
On 31 Mar 2009 13:12:10 GMT, Arno put finger to
keyboard and composed:

Franc Zabkar wrote:
On Tue, 31 Mar 2009 06:24:13 +1100, Franc Zabkar
put finger to keyboard and composed:

There are some claims on Steve Gibson's web site which I don't
understand.

Steve Gibson claims to be able to influence a HD's read AGC (automatic
gain control) circuit.

SpinRite's Defect Detection Magnetodynamics:
http://www.grc.com/srphysics.htm

He says that ..

"SpinRite is actually able to lower the amplification of the drive's
internal read-amplifier, then to cause the drive to encounter a
'minimum amplitude' data signal. If that data signal happened to fall
upon a spot of the drive's recording surface that is in any way 'weak'
then the drive will mis-read the data bit and return an error for the
sector containing the bad spot."

If I remember correctly, that claim was bogus to begin with.
What he could do with MFM/RLL is to test with a pattern, that
was more sensitve to surface problems, as it used the weak spots
of the modulation. Pretty much like black magic, so he possibly
dumbed it down. The actuall read amplifier gain is not accessible
and was not accessible in the MFM/RLL times as well.

AIUI, Steve Gibson does not claim to be able to control the read amp
gain directly. AGC relies on a rolling average, so Gibson "trains" the
read amp with strong signals, causing the AGC to reduce the gain, and
then he feeds the amp with weak signals while the gain is still low.

If I remember correctly, it is not an AGC, but rather the ganetizaton
strenght that can be influenced locally by writing the right data with
MFM. Does not work with todays modulation, though.

As for whether AGC was used in the early MFM days, I have a circuit
diagram for the 20MB drive that was used in the original IBM PC/AT.
The read amp appears to consist of two-stage differential op-amps with
fixed gains, ie no AGC. The print is hard to read, but the op-amps
look like NE592 video amps which are also specified for use as floppy
disc head amplifiers.

No surprise. I don't think you will have AGC in a HDD. Signals read
druing seeks would play havoc with it, IMO.

Here is a datasheet:
http://www.datasheetarchive.com/pdf-...DSA-463326.pdf

Well, no AGC on these. Just fixed gain.

However US patent #6839193, entitled "Method and apparatus for
determining read-to-write head offset of a disk drive", assigned to
Seagate, shows that the data field of each sector is preceded by an
"AGC data field which may be used to set read sensitivity".

The patent states that "when the read/write head is over the AGC data
field the gain of the recovered signal is adjusted to provide a
predetermined constant amplitude signal in order to reduce
demodulation errors."

So it appears to me that the read amp's AGC circuit is "trained"
during the "preamble" to the data field and is therefore not
influenced by the data pattern. Of course I'm assuming that the
technology described in the above patent (dated Jan 4, 2005) is
actually being implemented in current products.

No idea.

Arno

AIUI, the above patent talks about an AGC data field that is used to
train the servo amp so that it can accurately position the MR read
head. If a write is commanded, then the servo is jogged slightly
off-track to where the write coil is known to be. This may partly
explain why the seek time for writes is slightly greater than for
reads. Accurate positioning is dependent on accurate calibration of
the servo signal. For example, if an on-track read signal, after AGC,
is 10mV, then the write head's offset from the read head could be
where this signal drops to 8mV, say. AIUI these read-write head
offsets are calibrated at the factory and vary from track to track.

In addition to the above AGC data field, it appears that the user data
field has its own preamble of "training" bits which are laid down when
the data are written. In this case these training and sync bits are
used to condition the read amp when reading back the user data.

According to US patent number 6055118, entitled "holding automatic
gain control levels during read error recovery in a disc drive",
assigned to Seagate, the data block begins with a phase locked
oscillator (PLO) field which provides timing and amplitude information
for AGC and clocking. This is following by a training field and sync
field, the latter being used to identify the beginning of the user
data. The patent states that "the AGC first operates in an
'acquisition' mode during which large adjustments in the gain are made
in response to the initial signal levels provided by the preamp to
provide a coarse adjustment of the gain. The AGC then transitions into
a 'fine adjustment' mode, during which the AGC normally continues to
provide small, incremental adjustments to the gain as long as the read
channel is enabled as a particular sector is read".

So at first it appears that SpinRite may be able to influence the AGC
as claimed by Gibson. But then the patent goes on to describe what
happens when a read error is encountered. In this case the drive
re-reads the previous sector on the same track, and determines its
appropriate gain based on its PLO field. It then retries the bad
sector by turning off AGC and applying the *constant* gain derived
from the previous good sector. So it seems that any error that is a
consequence of an attempt to confuse AGC will be automatically
corrected by the drive on its second try, by *disabling* AGC.

BTW, the second patent applies to drives using PRML signal processing.

See http://www.google.com/patents?as_pnum=6055118

Well, doing that in firmware, with exact knowledge of the heads and
surfaces used, seems possible. SoinRite doing it seems to
be entirely marketing.

Arno

Barry OGrady wrote
Justin Goldberg wrote
Franc Zabkar wrote
Justin Goldberg wrote

Almost 60% of all hard drives damaged with bad sectors have an
incorrectly magnetized disk surface. We have developed an algorithm
which is used to repair damaged disk surfaces. This technology is
hardware independent, it supports many types of hard drives and
repairs damage that even low-level disk formatting cannot repair.

From following this group I've found that spinrite is absolute
rubbish, and now hdd regenerator is rubbish too?

I've tried to debate with the spinrite developers on news.grc.com
using some of the arguments against it from this newsgroup, but
it never goes anywhere, so I gave up:

I find that interesting because I have a 13G IDE hard disk that gave
read/write errors only on the top 3Gs. Scandisk showed seemingly
random bad sectors but Spinwrite did not find any problems.

Back in the AT days I had a full height 20MB MFM hard drive
that would show several bad sectors which Spinwrite would
fix until they showed up again maybe 3 months later.

Thats the well known sector drift problem seen with stepper motor head actuators.

The problem is that the position of the head varys with the temperature of the
drive so the sectors end up being scattered around the average position of the
head on that track. A reformat will fix that problem too because it write all the
sectors again. No longer a problem with drives that dont use stepper motors
to position the heads.

I don't seem to have the original post coming through. But this looks
promising:

http://www.cgsecurity.org/wiki/TestDisk
--
Ed Light

Better World News TV Channel:
http://realnews.com

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Iraq Veterans Against the War:
http://ivaw.org
http://couragetoresist.org

Send spam to the FTC at

Thanks, robots.