8 Nov
2007
8 Nov
'07
7:08 p.m.
I've got a Magnetic Peripherals 94211-91 SCSI drive here where the magnets for
the head stack seem to have been coated - at least around the edges - with
some form of paint. The paint's now begun to flake off, jamming up the head
assembly and preventing the drive from seeking properly (sometimes it wouldn't
spin up at all, other times it'd spin up, flop the heads around as much as it
could, and then power down again)
Anyone seen this particular mode of failure before? It's a new one on me...
With nothing to lose (after ruling out other possibilities, and I'd already
popped the lid on the drive expecting to see the aftermath of a head crash) I
removed the top magnet from the head assembly (without disturbing the heads
themselves or their pivot point) and cleaned everything out as best I could.
I was doing it more for giggles than anything - but I was amazed when, after
replacing the magnet, the drive actually powered up, stayed powered up, and
was detected by the OS.
It's not in particularly good health - I'm about 14MB into the archive process
and I've got about 500KB of dead blocks so far. Doubtful I'll be able to get
any meaningful data out of whatever filesystems were on it with that level of
corruption, but I suppose it'd be silly not to at least try!
cheers
Jules
8 Nov
8 Nov
6:10 p.m.
I've got a Magnetic Peripherals 94211-91 SCSI drive here where the magnets for
the head stack seem to have been coated - at least around the edges - with
some form of paint. The paint's now begun to flake off, jamming up the head
Snip:
Anyone seen this particular mode of failure before?
It's a new one on me...
Jules
This is unfortune fact of life with many rare-earth magnets, they
were sintered material and corrodes badly unless coated or plated
(chromed or nickel plated), some had poorer coating/plating and
begin to corrode and swells eventually burst maget's seams open and
get crap all over inside the hard drive.
:P
Still is even in modern hard drives when I had chance to take apart
and I had one not too old camrecorder (7 years old) die this way,
plated washer shaped magnet rotor burst open and jammed for the
capstan drive. Cost was too high even in parts alone and age of this
clinched it.
Cheers, Wizard
9 Nov
9 Nov
5:07 a.m.
jpero at sympatico.ca wrote:
Interesting... it's not something I've ever seen happen before - logic faults,
head crashes, spindle motor failures etc, yes, but it's the first time I've
seen this particular failure. I didn't even realise until I saw this drive
that drive magnets are coated in anything at all.
My archive attempt got 58MB into the 80MB drive (with about 2MB of broken
sectors) then the drive suddenly span down - I wonder if it maintains its own
error count on board and just declared itself broken. I'm just trying another
attempt now and if it's worth doing I'll merge the resulting two files and see
if I can get any meaningful data out of them.
The disk was from Torch's offices, and does seem to have an install of their
flavour of UNIX on (I'd wondered if it'd been wiped) - there's a chance
it's
got something interesting on it. I've managed to mount Torch partitions before
under Linux (using one of the UFS variants I think) so I'll give that a try -
it's just a question of whether the image is way too full of holes for the
filesystem driver to make any sense of it...
cheers
Jules
I've got a
Magnetic Peripherals 94211-91 SCSI drive here where the magnets for
the head stack seem to have been coated - at least around the edges - with
some form of paint. The paint's now begun to flake off, jamming up the head
Snip:
Anyone seen this particular mode of failure
before? It's a new one on me...
Jules
This is unfortune fact of life with many rare-earth magnets, they
were sintered material and corrodes badly unless coated or plated
(chromed or nickel plated), some had poorer coating/plating and
begin to corrode and swells eventually burst maget's seams open and
get crap all over inside the hard drive. 
4:38 a.m.
On Fri, 9 Nov 2007, Jules Richardson wrote:
I've got a Magnetic Peripherals 94211-91 SCSI
drive here where the magnets
for the head stack seem to have been coated - at least around the edges -
with some form of paint. The paint's now begun to flake off, jamming up the
head assembly and preventing the drive from seeking properly (sometimes it
wouldn't spin up at all, other times it'd spin up, flop the heads around as
much as it could, and then power down again)
Anyone seen this particular mode of failure before? It's a new one on me...
Oh yes... I have an HP 97536-60051 ESDI hard disk with the same symptoms.
The magnets are painted black and this paint flaked off (into lots of tiny
bits and also into dust). But the paint seems to be magnetic as the flakes
and (most of) the dust stick on the magnets and are hard to remove (even
with a brush and vacuum cleaner).
Christian
5:10 a.m.
Christian Corti wrote:
Exactly the same with my drive. Cleaning the top magnet wasn't so bad as I
could take it to the sink to do it and really work at it, but I couldn't do
much for the lower one - and obviously removing the whole head assembly is a
bad idea ;)
I don't think it's a failure mode I've ever seen mentioned on the list before.
I wonder it's it's one that's going to become more and more common as the
drives age, though? :-(
Anyone seen
this particular mode of failure before? It's a new one on
me...
Oh yes... I have an HP 97536-60051 ESDI hard disk with the same
symptoms. The magnets are painted black and this paint flaked off (into
lots of tiny bits and also into dust). But the paint seems to be
magnetic as the flakes and (most of) the dust stick on the magnets and
are hard to remove (even with a brush and vacuum cleaner). 
6:33 p.m.
Hi,
I don't think it's a failure mode I've
ever seen mentioned on
the list before. I wonder it's it's one that's going to become
more and more common as the drives age, though? :-(
I've never heard of this exact problem before either, but I do recall a
similar incident back in the early days of IDE.
Conner introduced a new line of drives, which after just a few months of use
started crashing en-masse. Seems they'd painted the arms on which the heads
were mounted, and said paint was flaking off....
TTFN - Pete.
10 Nov
10 Nov
10:08 p.m.
On Fri, 9 Nov 2007, Jules Richardson wrote:
Exactly the same with my drive. Cleaning the top
magnet wasn't so bad as
I could take it to the sink to do it and really work at it, but I
couldn't do much for the lower one - and obviously removing the whole
head assembly is a bad idea ;)
I don't think it's a failure mode I've ever seen mentioned on the list
before. I wonder it's it's one that's going to become more and more
common as the drives age, though? :-(
If you can get enough access to it, try a small length of steel wire or
maybe a small nail. It will pick up the magnetic junk that's stuck to the
magnet.
I've seen this sort of thing in high end VU meters used in the meter
bridges of large studio mixers before. Talk about a huge mess...
Some of the Simpson 260 meter movements had the same sort of problem,
though AFAIK they didn't use rare earth magnets.
11 Nov
11 Nov
10:51 a.m.
On Fri, 9 Nov 2007, Jules Richardson wrote:
> Exactly the same with my drive. Cleaning the top magnet wasn't so bad as
> I could take it to the sink to do it and really work at it, but I
> couldn't do much for the lower one - and obviously removing the whole
> head assembly is a bad idea ;)
Many hard drives allows you to swing the bottom magnet out without
moving the arm like Fujistu, Conner design of chassis. For ones you
need some kind of tool to gently ramp heads off the platters and
unbolted.
Somewhat OT but on topic about crumbling rare earth magnets:
While on the bursting rare earth magnets, I have got a set (under 3
years old) with a color wheel one of magnetic bearing suspension
went powdery (burst the plated protective shell) and gummed up the
works rendering the DLP's color wheel useless (siezed). Costs
between 150-180 for a colorwheel and labor not included. OUCH!
Makers of DLP sets went to ceramic air bearing instead for color
wheels. And can be installed in older ones and change a jumper and a
twreak to the color wheel index timing in some makes. Color wheels
is is 6 to 7 segments made of seperate pieces of thin optical glass
and plated with different colors (two sets of red, green and blue
usually) clamped to the aluminum/brass hub. Had to be precise made
and balanced to archive 9,000+ rpm and make little noise as possible.
Cheers, Wizard
6099
days inactive
6101
days old
7 comments
5 participants
participants (5)
-
cc＠corti-net.de -
classiccmp＠memory-alpha.org.uk -
jpero＠sympatico.ca -
julesrichardsonuk＠yahoo.co.uk -
tothwolf＠concentric.net