23 Oct
2015
23 Oct
'15
12:04 a.m.
Hi all --
I acquired a Symbolics 3640 today and it came equipped with two "large"
capacity Maxtor MFM drives (an XT-1140 and an XT-2190). The 1140 spins
up fine and we were able to image it using Dave Gesswein's MFM emulator
(yay).
The 2190 does not, and it fails in precisely the same way I've
personally seen three or four other Maxtor drives of the same era fail:
It spins up fine, but when it goes to load the heads, it sounds like the
voice coil positioner for the heads is "screaming" -- it emits a
high-pitched, quite loud whine/buzz which persists until you power the
drive down. The drive is unresponsive during this time.
I'm fairly sure it's not a head crash or anything like that; having gone
through this a year or so ago with a similar drive that was scratch
anyway, I opened it up and verified that the heads weren't stuck, and I
see no evidence of a head crash after disassembly.
Further, the fault does not appear to be on the logic board -- we
swapped in a board from a working 2190 tonight and afterwards the drive
exhibited the same symptoms.
I've had this happen to other 2190s and 1140s and a few of the ESDI
drives in the same family, some of which were working in my possession
for weeks before failing -- has anyone else seen this? Any ideas? I'd
kind of like to recover the data off of the 2190 from the 3640... drat.
Thanks,
Josh
23 Oct
23 Oct
1:19 a.m.
There is a plastic bumper in the head/disk assembly that turns to goo.
When the head retracts it hits the bumper and gets stuck in the goo. The goo will
eventually win. The head will no longer load. I can't say For sure this is your disk
problem but it was a verry common Maxtor failure.
Joe
On Oct 23, 2015, at 3:04 AM, Josh Dersch <derschjo
at gmail.com> wrote:
Hi all --
I acquired a Symbolics 3640 today and it came equipped with two "large"
capacity Maxtor MFM drives (an XT-1140 and an XT-2190). The 1140 spins up fine and we
were able to image it using Dave Gesswein's MFM emulator (yay).
The 2190 does not, and it fails in precisely the same way I've personally seen three
or four other Maxtor drives of the same era fail: It spins up fine, but when it goes to
load the heads, it sounds like the voice coil positioner for the heads is
"screaming" -- it emits a high-pitched, quite loud whine/buzz which persists
until you power the drive down. The drive is unresponsive during this time.
I'm fairly sure it's not a head crash or anything like that; having gone through
this a year or so ago with a similar drive that was scratch anyway, I opened it up and
verified that the heads weren't stuck, and I see no evidence of a head crash after
disassembly.
Further, the fault does not appear to be on the logic board -- we swapped in a board from
a working 2190 tonight and afterwards the drive exhibited the same symptoms.
I've had this happen to other 2190s and 1140s and a few of the ESDI drives in the
same family, some of which were working in my possession for weeks before failing -- has
anyone else seen this? Any ideas? I'd kind of like to recover the data off of the
2190 from the 3640... drat.
Thanks,
Josh
1:26 a.m.
On 10/23/15 1:19 AM, Joseph Lang wrote:
There is a plastic bumper in the head/disk assembly
that turns to goo.
When the head retracts it hits the bumper and gets stuck in the goo. The goo will
eventually win. The head will no longer load. I can't say For sure this is your disk
problem but it was a verry common Maxtor failure.
I thought that was true for Micropolis drives (like the DEC RD53, a
Micropolis 1325/1335) -- or does the Maxtor have a rubber bumper as
well? As I said I had one open and the heads were not stuck (I could
move them with my finger -- while the drive was spinning of course :))
and I didn't notice any goo-laden parts, but maybe I wasn't looking in
the right place...
- Josh
Joe
> On Oct 23, 2015, at 3:04 AM, Josh Dersch <derschjo at gmail.com> wrote:
>
> Hi all --
>
> I acquired a Symbolics 3640 today and it came equipped with two "large"
capacity Maxtor MFM drives (an XT-1140 and an XT-2190). The 1140 spins up fine and we
were able to image it using Dave Gesswein's MFM emulator (yay).
>
> The 2190 does not, and it fails in precisely the same way I've personally seen
three or four other Maxtor drives of the same era fail: It spins up fine, but when it
goes to load the heads, it sounds like the voice coil positioner for the heads is
"screaming" -- it emits a high-pitched, quite loud whine/buzz which persists
until you power the drive down. The drive is unresponsive during this time.
>
> I'm fairly sure it's not a head crash or anything like that; having gone
through this a year or so ago with a similar drive that was scratch anyway, I opened it up
and verified that the heads weren't stuck, and I see no evidence of a head crash after
disassembly.
>
> Further, the fault does not appear to be on the logic board -- we swapped in a board
from a working 2190 tonight and afterwards the drive exhibited the same symptoms.
>
> I've had this happen to other 2190s and 1140s and a few of the ESDI drives in the
same family, some of which were working in my possession for weeks before failing -- has
anyone else seen this? Any ideas? I'd kind of like to recover the data off of the
2190 from the 3640... drat.
>
> Thanks,
> Josh
>
>
1:33 a.m.
If the bumper is there it will be on the side wall of the HDA where the head actuator
would touch when retracted.
If the heads move freely you have a driver failure. the scream is the stepper motor trying
to move with only one phase working. (Also a common drive failure.)
Joe
On Oct 23, 2015, at 4:26 AM, Josh Dersch <derschjo
at gmail.com> wrote:
On 10/23/15 1:19 AM, Joseph Lang wrote:
There is a plastic bumper in the head/disk assembly that turns to goo.
When the head retracts it hits the bumper and gets stuck in the goo. The goo will
eventually win. The head will no longer load. I can't say For sure this is your disk
problem but it was a verry common Maxtor failure.
I thought that was true for Micropolis drives (like the DEC RD53, a Micropolis 1325/1335)
-- or does the Maxtor have a rubber bumper as well? As I said I had one open and the
heads were not stuck (I could move them with my finger -- while the drive was spinning of
course :)) and I didn't notice any goo-laden parts, but maybe I wasn't looking in
the right place...
- Josh
Joe
> On Oct 23, 2015, at 3:04 AM, Josh Dersch <derschjo at gmail.com> wrote:
>
> Hi all --
>
> I acquired a Symbolics 3640 today and it came equipped with two "large"
capacity Maxtor MFM drives (an XT-1140 and an XT-2190). The 1140 spins up fine and we
were able to image it using Dave Gesswein's MFM emulator (yay).
>
> The 2190 does not, and it fails in precisely the same way I've personally seen
three or four other Maxtor drives of the same era fail: It spins up fine, but when it
goes to load the heads, it sounds like the voice coil positioner for the heads is
"screaming" -- it emits a high-pitched, quite loud whine/buzz which persists
until you power the drive down. The drive is unresponsive during this time.
>
> I'm fairly sure it's not a head crash or anything like that; having gone
through this a year or so ago with a similar drive that was scratch anyway, I opened it up
and verified that the heads weren't stuck, and I see no evidence of a head crash after
disassembly.
>
> Further, the fault does not appear to be on the logic board -- we swapped in a board
from a working 2190 tonight and afterwards the drive exhibited the same symptoms.
>
> I've had this happen to other 2190s and 1140s and a few of the ESDI drives in the
same family, some of which were working in my possession for weeks before failing -- has
anyone else seen this? Any ideas? I'd kind of like to recover the data off of the
2190 from the 3640... drat.
>
> Thanks,
> Josh
10:16 a.m.
On 10/23/2015 09:10 AM, Al Kossow wrote:
Maxtor drives have a very distinctive (and loud) recal
sound.
And some Maxtor drives are Miniscribes. I have one (I think it's an
ESDI, IIRC) that has a beautiful Miniscribe HDA casting, with a Maxtor
label wrapped around it.
Ah, the good old days of brick-filled boxes...
--Chuck
12:48 p.m.
New subject: Miniscribe "bricks" (was Re: Common Maxtor MFM drive failure mode -- any ideas?)
----- Urspr?ngliche Message -----
Haha, I guess you're alluding to the massive scam with the bricks, Miniscribe did back
them to pretend stocks full of disk drives...
BTW, are there any other similar stories from the disk drive buisiness back in these
days?
Pierre
Von: Chuck Guzis <cclist at sydex.com>
An: General Discussion: On-Topic and Off-Topic Posts <cctalk at classiccmp.org>
CC:
Gesendet: 18:16 Freitag, 23.Oktober 2015
Betreff: Re: Common Maxtor MFM drive failure mode -- any ideas?
On 10/23/2015 09:10 AM, Al Kossow wrote:
Maxtor drives have a very distinctive (and loud)
recal sound.
And some Maxtor drives are Miniscribes. I have one (I think it's an
ESDI, IIRC) that has a beautiful Miniscribe HDA casting, with a Maxtor
label wrapped around it.
Ah, the good old days of brick-filled boxes...
--Chuck
1:35 a.m.
Sorry I didn't notice you said voice coil. The stepper comment is moot.
Joe
On Oct 23, 2015, at 4:26 AM, Josh Dersch <derschjo
at gmail.com> wrote:
On 10/23/15 1:19 AM, Joseph Lang wrote:
There is a plastic bumper in the head/disk assembly that turns to goo.
When the head retracts it hits the bumper and gets stuck in the goo. The goo will
eventually win. The head will no longer load. I can't say For sure this is your disk
problem but it was a verry common Maxtor failure.
I thought that was true for Micropolis drives (like the DEC RD53, a Micropolis 1325/1335)
-- or does the Maxtor have a rubber bumper as well? As I said I had one open and the
heads were not stuck (I could move them with my finger -- while the drive was spinning of
course :)) and I didn't notice any goo-laden parts, but maybe I wasn't looking in
the right place...
- Josh
Joe
> On Oct 23, 2015, at 3:04 AM, Josh Dersch <derschjo at gmail.com> wrote:
>
> Hi all --
>
> I acquired a Symbolics 3640 today and it came equipped with two "large"
capacity Maxtor MFM drives (an XT-1140 and an XT-2190). The 1140 spins up fine and we
were able to image it using Dave Gesswein's MFM emulator (yay).
>
> The 2190 does not, and it fails in precisely the same way I've personally seen
three or four other Maxtor drives of the same era fail: It spins up fine, but when it
goes to load the heads, it sounds like the voice coil positioner for the heads is
"screaming" -- it emits a high-pitched, quite loud whine/buzz which persists
until you power the drive down. The drive is unresponsive during this time.
>
> I'm fairly sure it's not a head crash or anything like that; having gone
through this a year or so ago with a similar drive that was scratch anyway, I opened it up
and verified that the heads weren't stuck, and I see no evidence of a head crash after
disassembly.
>
> Further, the fault does not appear to be on the logic board -- we swapped in a board
from a working 2190 tonight and afterwards the drive exhibited the same symptoms.
>
> I've had this happen to other 2190s and 1140s and a few of the ESDI drives in the
same family, some of which were working in my possession for weeks before failing -- has
anyone else seen this? Any ideas? I'd kind of like to recover the data off of the
2190 from the 3640... drat.
>
> Thanks,
> Josh
9:08 a.m.
On 10/23/15 12:04 AM, Josh Dersch wrote:
The 2190 does not, and it fails in precisely the same
way I've personally seen three or four other Maxtor drives of the same era fail: It
spins up fine, but when it goes to load the heads, it sounds
like the voice coil positioner for the heads is "screaming" -- it emits a
high-pitched, quite loud whine/buzz which persists until you power the drive down. The
drive is unresponsive during this time.
Sounds like a failure of the positioner servo. It is on a separate platter.
It should be possible to build something to open the servo loop and position the heads
externally. One of those projects that I would like to get to some day. I have about 50
Maxtor ESDI drives from Apollo's build system that I REALLY want to get the data off
of
and I suspect there will be the same sort of problem with them.
9:31 a.m.
The 2190 does not, and it fails in precisely the
same way I've personally seen three or four
other Maxtor drives of the same era fail: It spins up fine, but when it goes to load the
heads,
it sounds like the voice coil positioner for the heads is "screaming" -- it
emits a high-pitched,
quite loud whine/buzz which persists until you power the drive down. The drive is
unresponsive
during this time.
Sounds like a failure of the positioner servo. It is on a separate platter.
I wonder if there is a separate velocity feedback coil (not unuheard-of, but not that
common on
smaller dries)? If so, a problem with that would cause the servo to go unstable.
Normally a total loss of servo information causes the heads to hit one or other of the end
stops,
but there are many other posibilities.
It should be possible to build something to open the
servo loop and position the heads
externally. One of those projects that I would like to get to some day. I have about 50
Remember that in the end all you can control is the current through the 'motor'
coil (not
strictly a voice coil in this drive, but the same idea), and that applies a force
(==acceleration)
to the heads. Normally this is a closed loop system, signals picked up by the servo heads
tell the control electronics of the heads are off-track (and basically how far) this is
used
to generate a correction signal to feed to the coil.
If you open the servo loop, you will be able to accelerate the heads but little more. You
will
not AFAIK be able to keep them on track.
Of course the other issue is that if the servo head/platter/preamp is malfunctioning, will
the
data heads be any better? You may not be able to read anything much even if you are
on-track.
-tony
12:39 p.m.
On Fri, Oct 23, 2015 at 9:08 AM, Al Kossow <aek at bitsavers.org> wrote:
On 10/23/15 12:04 AM, Josh Dersch wrote:
The 2190 does not, and it fails in precisely the same way I've personally
Hm. It would be interesting to test the servo failure hypothesis; I don't
suppose anyone has a service manual for these things so I know what stuff
to probe? (Nothing on Bitsavers and a casual Google search turns up
nothing of interest.)
- Josh
seen three or four other Maxtor drives of the
same era fail: It spins up
fine, but when it goes to load the heads, it sounds
like the voice coil positioner for the heads is "screaming" -- it emits a
high-pitched, quite loud whine/buzz which persists until you power the
drive down. The drive is unresponsive during this time.
Sounds like a failure of the positioner servo. It is on a separate platter.
It should be possible to build something to open the servo loop and
position the heads
externally. One of those projects that I would like to get to some day. I
have about 50
Maxtor ESDI drives from Apollo's build system that I REALLY want to get
the data off of
and I suspect there will be the same sort of problem with them.
2:43 p.m.
On 10/23/15 12:39 PM, Josh Dersch wrote:
I don't
suppose anyone has a service manual for these things so I know what stuff
to probe? (Nothing on Bitsavers and a casual Google search turns up
nothing of interest.)
Service manuals/schematics/ASIC info is EXTREMELY difficult to get for anything beyond
the stepper motor era. What is on bitsavers are the results of my searching for
20+ years, and that isn't much. I've been digging into early 90's DC2000 QIC
drives over the
past couple weeks, and hw docs for them are just as difficult to find.
24 Oct
24 Oct
10:15 a.m.
On 10/23/15 2:43 PM, Al Kossow wrote:
On 10/23/15 12:39 PM, Josh Dersch wrote:
That's pretty much what I figured. I took a closer look at one of the
other dead XT2190s I have that I'd opened up to inspect awhile back and
there are a few ICs surface-mounted to the flat ribbon cable running to
the head assembly. I suppose it's likely that one of these has failed,
though actually repairing it would be a trick involving some very
careful disassembly in a very clean environment. (And a nonexistent
service manual.)
- Josh
I don't
suppose anyone has a service manual for these things so I know what
stuff
to probe? (Nothing on Bitsavers and a casual Google search turns up
nothing of interest.)
Service manuals/schematics/ASIC info is EXTREMELY difficult to get for
anything beyond
the stepper motor era. What is on bitsavers are the results of my
searching for
20+ years, and that isn't much. I've been digging into early 90's
DC2000 QIC drives over the
past couple weeks, and hw docs for them are just as difficult to find.
11:40 a.m.
That's pretty much what I figured. I took a closer look at one of the
other dead XT2190s I have that I'd opened up to inspect awhile back and
there are a few ICs surface-mounted to the flat ribbon cable running to
the head assembly. I suppose it's likely that one of these has failed,
though actually repairing it would be a trick involving some very
careful disassembly in a very clean environment. (And a nonexistent
service manual.)
I have had (in other devices) dry joints on SMD devices on flexiprints. But
resoldering them inside the HDA is not going to be easy...
Most likely those ICs are head switch/preamp devices and the servo head
preamplifier. They are very likely to be custom.
On older/larger drives (the sort of thing I am more likely to work on) the ICs were
often DIL packages on a normal PCB. Often you couldn't replace them without opening
the HDA :-(. Micropolis had a nice feature on the 1200 series (8" hard drives) though
--
the PCB was mounted over a hole in the HDA casing (obviously with a gasket). The heads
were wired to the inside face of the PCB, the cable to the logic board plugged into the
outside face. The ICs were plugged into turned pin sockets on the outside. So on that
drive you could field-replace them. They were custom chips, though. And of course you
couldn't replace soldered parts, like the decoupling capacitors as the solder joints
formed
part of the HDA seal...
-tony
- Josh
2:22 p.m.
On 10/24/15 11:40 AM, tony duell wrote:
Most likely those ICs are head switch/preamp devices
and the servo head
preamplifier. They are very likely to be custom.
Silicon Systems was a common supplier in the 80s to mid-90s, which is why their
Storage Products data books have been scanned.
25 Oct
25 Oct
1:48 p.m.
On Oct 24, 2015, at 14:22, Al Kossow <aek at
bitsavers.org> wrote:
On 10/24/15 11:40 AM, tony duell wrote:
Just thought I'd mention that my first "real" job was at Silicon Systems.
Started as an applications engineer for motor drivers, then moved into chip design for
interface controllers. I don't think I have any old databooks any more. Glad to hear
that they're being preserved.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
Most likely those ICs are head switch/preamp
devices and the servo head
preamplifier. They are very likely to be custom.
Silicon Systems was a common supplier in the 80s to mid-90s, which is why their
Storage Products data books have been scanned.
24 Oct
24 Oct
4:36 p.m.
On 10/24/15 11:40 AM, tony duell wrote:
So, a quick update here (and some idiocy on my part):
I (carefully) opened up the Symbolics' XT-2190 and took a quick look;
the rubber bumper that Joseph mentioned (and it is rubber in this one --
my other opened drive has a plastic bumper) has started
crumbling/turning to goo. Portions of it had already chipped off. This
bumper acts as both the start and end stop for the head assembly.
I removed the remainder of it (it's held on with a circlip) and spun the
drive up (to see if they had just been stuck to the bumper) to the same
effect as before -- the heads recal and then it just buzzes. Then I
gave the heads a nudge just to see what would happen, and... they ran
off the end of the platters (no stop anymore) and well, I feel kinda
stupid. Sigh.
To add insult to injury, one of the heads is loose (the glue holding it
on dried up and it fell off after the impact of running off the platter)
so this drive is basically toast. At least now I can kind of see how
one takes this drive apart to remove the spindle; if I get overly
ambitious and find a working sacrificial XT2190 to start with I could
almost see myself doing a spindle replacement surgery to see if I can
recover the data.
I think I'll stick to solid state devices for awhile.
- Josh
That's
pretty much what I figured. I took a closer look at one of the
other dead XT2190s I have that I'd opened up to inspect awhile back and
there are a few ICs surface-mounted to the flat ribbon cable running to
the head assembly. I suppose it's likely that one of these has failed,
though actually repairing it would be a trick involving some very
careful disassembly in a very clean environment. (And a nonexistent
service manual.)
I have had (in other devices) dry joints on SMD devices on
flexiprints. But
resoldering them inside the HDA is not going to be easy...
Most likely those ICs are head switch/preamp devices and the servo head
preamplifier. They are very likely to be custom.
On older/larger drives (the sort of thing I am more likely to work on) the ICs were
often DIL packages on a normal PCB. Often you couldn't replace them without opening
the HDA :-(. Micropolis had a nice feature on the 1200 series (8" hard drives)
though --
the PCB was mounted over a hole in the HDA casing (obviously with a gasket). The heads
were wired to the inside face of the PCB, the cable to the logic board plugged into the
outside face. The ICs were plugged into turned pin sockets on the outside. So on that
drive you could field-replace them. They were custom chips, though. And of course you
couldn't replace soldered parts, like the decoupling capacitors as the solder joints
formed
part of the HDA seal...
-tony
- Josh
25 Oct
25 Oct
4:56 a.m.
On Oct 24, 2015 6:36 PM, "Josh Dersch" <derschjo at gmail.com> wrote:
I've successfully glued heads back on and have cleaned platters of goo
before. I had a full height 40 MB CDC drive from a PC AT if I recall that
wouldn't start reliably, but once it was spinning, worked fine. I took it
apart in a clean room and had the same issue as you with heads falling off.
I reattached them with clear fingernail polish for fear that superglue
would leave nasty cyanoacrylate fumes everywhere.
The drive worked fine for several months after the repair and a low level
format. It suffered the same problem again and I opted not to fix it a
second time.
It took me a while to figure out an optimal way to put the head back
together. I ended up separating the heads with the big ends of zip ties
which happened to be just the right height. Once I got everything mounted
back in, I just pushed it over the platters, causing the zip tie pieces to
be pushed out, and the heads to fall back over the platters.
Ahh, found the album (which I think I've shared before):
http://imgur.com/a/LbfZb
Kyle
To add insult to injury, one of the heads is loose (the glue holding it
on dried up
and it fell off after the impact of running off the platter) so
this drive is basically toast. At least now I can kind of see how one
takes this drive apart to remove the spindle; if I get overly ambitious and
find a working sacrificial XT2190 to start with I could almost see myself
doing a spindle replacement surgery to see if I can recover the data.
3303
days inactive
3305
days old
18 comments
8 participants
participants (8)
-
aek＠bitsavers.org -
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
derschjo＠gmail.com -
joe.lang.0000＠gmail.com -
kylevowen＠gmail.com -
nf6x＠nf6x.net -
p.gebhardt＠ymail.com