Very Computer

My workstation (Super Workstation X/40) boots up using a Quantum Empire
1080S drive. Recently, it gets up to the following line and freezes:

and args:

That's as far as it will go. I called someone on site here about it and she
said the drive may be sticking, due to some gel substance (put there by the
drive manufacturer) that has coagulated inside the drive. She actually
suggested hitting the drive a few times (not TOO hard, of course) to loosen
it.

Has anyone else had this happen? Or have has anyone heard of this
phenomenon?

Alan McCoy

|
| My workstation (Super Workstation X/40) boots up using a Quantum Empire
| 1080S drive. Recently, it gets up to the following line and freezes:
|

| and args:
|
| That's as far as it will go. I called someone on site here about it and she
| said the drive may be sticking, due to some gel substance (put there by the
| drive manufacturer) that has coagulated inside the drive. She actually
| suggested hitting the drive a few times (not TOO hard, of course) to loosen
| it.
|
| Has anyone else had this happen? Or have has anyone heard of this
| phenomenon?

Yes, the phenomenon is termed "stiction" (aka static friction).

At one time both Seagate and Quantum were notorious for their drives not
spinning-up during power-on because the heads would be stuck in the park
position due to having over-lubricating the platters during manufacture.

The operative word here is "meniscus" (of the lubricant with the parked
heads).

I have about 20MB of text files describing this (and related) problems with
Seagate drives gleaned from several score Usenet newsgroups between 1985 and
1995.  The only recourse is to get the drive powered-up and spinning, then
move your data off ASAP or never let the drive power-down once it's spinning;
it's no longer cost-effective to replatter a drive given the low cost of HDs
today.

Summarizing how the problem starts: workers in the drive assembly plants
were supposed to wipe a lubricant-laden cloth ONCE on each platter, then
pass the platters onwards for final optical inspection.  The workers would
be paid by how many platters passed the inspection.  Oftimes the cloths
were not properly treated (with lubricant) or the workers would sloppily
wipe the platter, causing it to be rejected during optical inspection (and
thus the workers received less money).  To avoid a pay reduction, they
would both oversaturate the cloths and wipe the platters more than once.

Since the optical inspection was with a microscope, the inspector really
couldn't determine whether the platter had the correct mono-molecular layer
(of the highly proprietary lubricants (each HD mfr has their own formulation))
and so would "pass" a platter that might have 2 to 4 times the design level
of lubricant because the surface appeared "smooth and flat".

Later on, after a drive has been in service, the lubricant would tend to
accumulate in 2 areas: near the outer edge (cylinder 0), and near the center
(the parking zone).  If you open such a drive, the accumulations are readily
apparent under bright and/or sun light (yes, I've done this).

When you power-down a drive whose platters were overlubricated, the heads
move to the center and then rest.  Now a meniscus forms, and this holds the
heads in place and prevents the platters from spinning when the drive is
powered-up (akin to how a car's disk brake clamps onto a disk to reduce/stop
velocity).

Your colleague's suggestion to "whack the drive" is, in fact, a temporary
solution (to break the meniscus) and permit the drive to spin up and let
you get the data off.  I once had a Seagate ST157N that required turning
the main spindle with a pair of pliers to break the stiction and get the
drive started.

FWIW, my experiences with the Quantum Empire HD series are dismal.

Also FWIW, any HD manufacturer who claims 100,000+ hours MTBF has mush for
brains and is lying.  And don't listen to any BS about statistics and
sampling and other such poppycock for as Samuel Clemens (aka Mark Twain)
once is said to have commented (paraphrased): "There are lies, damned lies,
and then there are statistics."  :-)

Use some good (not common) sense and figure it out for yourself: 24 hours
in a day and 365.2425 days in a year = 8765 hours in a year.  A so-called
100,000 hour MTBF (Mean Time Between Failure) is 11.5 years.  Yeah, sure,
get real, how many HDs last that long in actual service?

Any rotating, heat-producing mechanical kluge such as a disk drive WILL fail;
the only unknown is when.

My own experience with thousands of HDs suggests that after 4 years of
continuous service one better have a replacement/spare handy and good
backups.  With today's HD prices it's prudent to have 2 spare HDs on-site
for every 5-10 workstations in your shop unless you can afford prolonged
downtime until replacements are purchased, shipped, received and installed.

Thad

Thanks for the help and advice on this, everyone.  I finally got  the drive
to work (with a little jiggling and a couple of taps.)

Alan McCoy

This was one reason the 1080S disappeared from the market
earlier than it's cousins.

Quote:>Yes, the phenomenon is termed "stiction" (aka static friction).
>At one time both Seagate and Quantum were notorious for their drives not
>spinning-up during power-on because the heads would be stuck in the park
>position due to having over-lubricating the platters during manufacture.

Quote:>The operative word here is "meniscus" (of the lubricant with the parked
>heads).
>I have about 20MB of text files describing this (and related) problems with
>Seagate drives gleaned from several score Usenet newsgroups between 1985 and
>1995.  The only recourse is to get the drive powered-up and spinning, then
>move your data off ASAP or never let the drive power-down once it's spinning;
>it's no longer cost-effective to replatter a drive given the low cost of HDs
>today.

Quote:>Summarizing how the problem starts: workers in the drive assembly plants
>were supposed to wipe a lubricant-laden cloth ONCE on each platter, then
>pass the platters onwards for final optical inspection.  The workers would
>be paid by how many platters passed the inspection.  Oftimes the cloths
>were not properly treated (with lubricant) or the workers would sloppily
>wipe the platter, causing it to be rejected during optical inspection (and
>thus the workers received less money).  To avoid a pay reduction, they
>would both oversaturate the cloths and wipe the platters more than once.
....
>Later on, after a drive has been in service, the lubricant would tend to
>accumulate in 2 areas: near the outer edge (cylinder 0), and near the center
>(the parking zone).  If you open such a drive, the accumulations are readily
>apparent under bright and/or sun light (yes, I've done this).
>When you power-down a drive whose platters were overlubricated, the heads
>move to the center and then rest.  Now a meniscus forms, and this holds the
>heads in place and prevents the platters from spinning when the drive is
>powered-up (akin to how a car's disk brake clamps onto a disk to reduce/stop
>velocity).

However 'stiction' is the term applied to the phenomenon that
occurs when you place two highly polished surfaces in contact
with each other.   These surfaces are so polished that air can
not leak between them and air pressue will hold them together.

In the machine industry there are meauring blocks called 'gage
blocks'.  Piece of metal accurately calibrated to specific
size.    If you take two of these and put them together with a
twisting motion - called wringing - the two will stay together
as the air-pressure will keep them together.

I worked in a multi-track recording studio years ago - and used
hundreds of miles of tape.  (We were in the end column $50,000+
from Scotch, Ampex and Agfa).

I was talking with an FE from AGFA and they had the stiction
problem with some 2" media they were working on.

As you use a tape recorder the heads become more and more
polisdhed by the action of the tape.  AGFA's tape was highly
calendared (I never saw this but the early test samples I had
of what was to become Scotch 250 were so highly calendared that
you could almost use it as a mirror to shave with - something
I've avoided doing (shaving) for many many years).

The tape would be played and when it came to a stop the action
was like the wringing together of the gage blocks.  When it
came time to start the 'stiction' between the tape head and the
tape prevented it from moving.

Quote:>Your colleague's suggestion to "whack the drive" is, in fact, a temporary
>solution (to break the meniscus) and permit the drive to spin up and let
>you get the data off.  I once had a Seagate ST157N that required turning
>the main spindle with a pair of pliers to break the stiction and get the
>drive started.
>FWIW, my experiences with the Quantum Empire HD series are dismal.

Quote:>Use some good (not common) sense and figure it out for yourself: 24 hours
>in a day and 365.2425 days in a year = 8765 hours in a year.  A so-called
>100,000 hour MTBF (Mean Time Between Failure) is 11.5 years.  Yeah, sure,
>get real, how many HDs last that long in actual service?

Quote:>My own experience with thousands of HDs suggests that after 4 years of
>continuous service one better have a replacement/spare handy and good
>backups.  With today's HD prices it's prudent to have 2 spare HDs on-site
>for every 5-10 workstations in your shop unless you can afford prolonged
>downtime until replacements are purchased, shipped, received and installed.

It's been a long time since the days of the Unix PC from AT&T.

I just reached up on my shelf and found that I received the
WD2010 chips from you in August of 1989.  (For the uninformed
those replaced the  1010 ?? - so that we could use drives with
more than 8 heads and break the 80MB  limit on those old
machines).  (are you still in Los Altos?)

--