PDA

View Full Version : Death by stiction



eupeptic
August 15th, 2008, 03:18 AM
So, I'm going through the old hardware I have before I give it away to you guys, and the thing I've been working on lately is an HP Vectra LS/12 (http://hpmuseum.net/display_item.php?hw=220) laptop (same as the Zenith SuperSport 286) with a 40 MB hard drive (Conner CP-343). A few days ago I charged the battery pack (for about 18 hours), then I ran the laptop to see how long the battery would power it (IIRC, it went for 3.5 hours with the display on, hard drive spinning, and CPU at full speed - which really surprised me as I didn't expect it to last for more than about 30 minutes, considering how old it is). Since then I created the setup disc so I could actually use it (the CMOS battery is dead), but I haven't been able to get the hard drive working (it would either/both give a drive not ready error, or an error about the controller). Today I took it apart again to see if there was a problem with the connection to the hard drive (I'm not sure if there was or wasn't as it slipped right off when I lifted the hard drive - however, I'm quite certain that it was on right as I did hear the hard drive spinning when I did the battery test). When I reconnected the drive and powered it on it would make a few quick attempts to spin up (it may have been doing this prior to me taking it apart, though I wasn't able to hear it), then pause for a second, then attempt to spin up again, and repeat this process one or two more times without success. Knowing that stiction was the problem I unplugged the drive and rotated it in my hand a few times (along the axis of the spindle), then it made a light scraping noise that I thought was due to the head assembly moving about the surface of the discs rather than being parked (which I was not happy to hear, but I couldn't do anything about it). Then I plugged it back in, powered it up and this time it spun, but it also made a clicking noise on each rotation and some scraping noise, so I quickly shut it off. At that point I was thinking that I'm screwed. Not being able to think of anything else I could do, and being quite confident that running it like that would cause further damage to the drive I decided to take it apart. What I found was this, one of the heads broke off:

http://img93.imageshack.us/img93/6504/dsc04914no3.th.jpg (http://img93.imageshack.us/my.php?image=dsc04914no3.jpg)

http://img93.imageshack.us/img93/7672/dsc04912rx7.th.jpg (http://img93.imageshack.us/my.php?image=dsc04912rx7.jpg)

:( :(

There wasn't any important data on it as my dad bought this laptop in the mid '90s and he probably only used it for a few months (it's been sitting ever since, though I did power it up for a few minutes last September when I found the power supply for it), but it would have been nice to see what was on there (he only used it as a word processor, and there probably wasn't much else on it).

I do seem to recall running across a relatively new hard drive (less than a few years old) that wouldn't spin up (I think it was one of the first Seagate drives that had fluid-dynamic bearings), and rotating the body along the axis of the spindle got it working again. But that force can be too much to ask of the heads on an old 3.5" drive... (Next time I run into an old drive that's suffering from stiction I'll use a different technique - a slow acceleration with a quick stop [while holding the drive so that the quick stop should cause the spindle to rotate in its normal direction rather than back and fourth] - hopefully that would work better.)

About a month ago or so I did run across a site that had information about disabling one of the heads on a failed hard drive (I'm not certain, but I think it was for ST412/ST506 drives) so if I can find that again I'd like to look into doing that because 30 MB is better than no hard drive at all (although I don't know if the BIOS would support it - as it only gives you the options to use drives 1-50, and you can't use custom settings). The good news is that while looking at the platters it is possible to see where the head became stuck (2nd picture), and I don't see any of those marks for any of the other heads.

One question I have is once a hard drive shows the signs of stiction, might it be at all possible to improve its condition (for example, through certain usage patterns with the hope of decreasing the amount of lubricant that has accumulated on the head(s)), or does it only get worse with use? (Has anyone here been reluctant to give up on a drive that has failed due to stiction?)

(And if anyone noticed in the first picture, one of the screws is missing from the spindle motor, and it was like that when I took it apart - maybe that was their method of balancing the disk assembly, since it appears that the torx screws are the main ones... Also in that picture, the black plastic arm next to the head assembly easily moves out of the way when the broken head on the disk spins around in the normal direction - so that's what was making the clicking noise when I turned it on after attempting to unstick the heads.)

QuantumII
August 15th, 2008, 06:06 AM
Isn't it possible to get a compatible HDD off ebay or somewhere else ? It's a standard IDE drive (I have a SuperSport/286 myself)

Here's a webshop that sells them Recycled Goods Conner C343 (http://www.recycledgoods.com/itemdetails.aspx?productID=7405) but it's a bit much to pay $81 for it I think..

You could always get a donor drive, and if you are really brave, replace the broken head stack with the one from a working one :)

Terry Yager
August 15th, 2008, 08:31 AM
If you have no luck getting it working, I'll buy it off ya for parts (especially the battery pack).

--T

carlsson
August 15th, 2008, 10:48 AM
I presume the power supply is a different model than the Zenith Z-Sport 425S one?

Druid6900
August 15th, 2008, 11:04 AM
All my successful attempts to overcome "stiction" have been by turning the drive spindle side up, putting a drop of 3-in-1 penatrating oil on the spindle shaft and letting it sit there for a while (hours).

Then hooking it up to a standard PSU and turning it on. in the majority of cases, this is enough to take care of the problem and letting it run in this position allows for full coating.

Sometimes, GENTLY moving the motor shell a couple of degrees, back and forth is necessary and several power on/off cycles.

After running for a few hours, turning it off, another drop of oil before it cools and another couple of hours of spinning.

I have drives that I did a couple of years ago and then set aside and about a month ago, I gathered them up, powered them and they still spin right up.

QuantumII
August 16th, 2008, 12:33 AM
As far as I can understand, his problem was that the actual head stuck to the platter, thus preventing it from spinning up. when the drive finally spun up, it had torn the head off the head arm.

eupeptic
August 16th, 2008, 01:59 AM
QuantumII, I've been pondering about whether this would support an IDE drive or not. Although later motherboards natively support IDE drives (as yours does), if you look at the pictures for the CP-343 on the link you gave (which is something I hadn't come across while looking for information on the drive, despite how little information is available online) you'll see that it has a 20-pin interface (which includes power). That cable goes to a daughterboard (here's a picture of it (http://img382.imageshack.us/my.php?image=dsc04923ya7.jpg)) which plugs into your standard 40-pin IDE connector on the motherboard, and if one were to use an IDE drive on this laptop your only option would be to plug it into the connector that daughterboard is plugged into. But I figured it wouldn't be that simple - and today I found this discussion (http://groups.google.com/group/comp.sys.zenith/browse_thread/thread/106e3059e3459e/4d989563ca58311f) in comp.sys.zenith which describes a hardware modification that may be necessary. (Looking at the pictures I took of the motherboard, the first portion of that modification has already been done, but the second and third parts haven't.) Then there's the issue of BIOS support; the largest drive supported is 155 MB.

Terry, I'd have to play around with it some more to be certain that the battery pack can continue to do 3.5 hours. (After that test I ran the battery pack down to 1 volt with a flashlight [for a little over an hour I think]. During that time I read this FAQ (http://www.repairfaq.org/ELE/F_NiCd_Battery.html) about rechargeable batteries [which provides a good overview of how batteries work so you can understand why certain procedures should or should not be done] and it essentially states that the recommendations about the memory effect of NiCd batteries is often overstated, and you can damage your batteries [namely battery packs] by draining them significantly. After that I recharged it for about 8 hours [as is recommended in the manual], then when I ran the laptop again [just poking around with it, forgetting that I didn't have the power adapter plugged in] it gave the low battery warning after about 30 or 45 minutes [I wasn't keeping track of time, but it wasn't long]. It could just be that it requires more than 8 hours of charge, though I do have an AC power meter and the AC adapter was only using 0.1 amps at the end of that 8-hour charge [which is the same amount of current as when it's not plugged into anything]. It could also be due to the hard drive not spinning up [and maybe it was just sucking down the batteries], though I'm not certain.) Either way, though, if you have your own battery pack it should be possible to replace the batteries in it with new NiCd D-cells (for example (http://store.shorelinemarket.com/nidsi50mahre.html); although this is more expensive than I would have imagined - 10 batteries are used in the battery pack). But, we'll see how things pan out when I post a thread to give the laptop away. (I'd kind of prefer that it all go to one person, but if you guys are more interested in splitting it up then that's fine too).

carlsson, I don't know what the specifications are for the AC adapter for the Z-Sport 425S (a variety are available here (http://www.zdsparts.com/acadapt.htm)), but this is (http://www.recycledgoods.com/itemdetails.aspx?productID=13680) the exact one I have. (Model number 150-308, 2 amps DC at 16.5 volts, and positive on the outside and negative in the center.)

Druid6900, cool. My main interest is in how much I should be kicking myself for doing something I know I shouldn't have done (after posting my message I recalled that I have a Seagate ST-225 or ST-251 hard drive that when I used it last about 10 years ago wouldn't spin up one time after having used it for a while and letting it sit for a day or more, and I'm pretty certain that I rotated the drive in a manner which should cause most of the rotational force to spin the platters in their normal direction rather than back and fourth (which is a bit of a challenge to do with a 5.25" drive while being careful not to drop it) - as I had heard of others who broke the heads off while attempting to unstick them and I wanted to be careful to avoid doing that); I just was too eager to get the drive working again and wasn't thinking that the head would be stuck to the surface of the platter to such an extent that it could cause it to break off - instead, I was thinking about how it was working fine just a few days prior, and I wasn't aware that a significant amount of use [that 3.5 hour test] after a long period without use could end up resulting with the head stuck to the platter) - if I hadn't broken the head off and if it would have been possible to get the drive back to a usable state, then I'd really wish I hadn't done that. (It's not possible to get to the spindle motor without disassembling the drive.) One odd thing is that while this drive is supposed to park its heads automatically, the heads were sitting in the center when they got stuck...

And QuantumII, I actually think the head broke off while I was rotating the drive in my hand since there was nothing that I could tell going on during the first few rotations, then something happened and I heard what I presumed to be the head assembly moving back and fourth over the surface of the platters. The wires may have broken off when I powered it up, though...

Lastly, I believe that this is the software (http://www.acedre.com/pc3000win.html) that I heard that people can use to disable one of the heads on their hard drive so they can still get some use out of it - and I doubt it'd work with the drive I have (not to mention that it's much more expensive than simply getting a replacement drive).

Druid6900
August 16th, 2008, 02:12 PM
I can't say that, in all the years I've been repairing computers, that I ever came across an incidence of a head being "welded" to a platter.

Some drives, such as the ST-213, I believe, have this "cowling" like affair over the motor shaft end, but, it does unscrew. You just have to remove the logic board and immobilize the shaft while doing it. Others have to be pried off.

eupeptic
August 16th, 2008, 08:53 PM
Today I decided to remove the broken head from the platter so I can see if there's any possibility that I can use the drive with one less head, and I expected it to slide right off when I grabbed it with the tweezers, but it didn't. Seeing that it was being stubborn I decided to slowly increase the amount of force on the head and see if I can estimate how much force it takes to get it to come off. (If you look at the first picture in my first post I pretty much had the drive like that [though I was holding it my hand] and was trying to slide the head towards the outside edge of the platter.) After a few seconds of increasing the pressure on the head it finally popped off, and I'd guess that it took somewhere around 4 or 5 ounces of lateral force to get it to do that. :o Though I didn't test this, I'm certain that if I had it sitting on a flat surface the entire drive would slide around while pushing on the head. I don't think I would have been able to pick up the entire drive, though [either while it's sitting flat and lifting straight up on the head, or sitting on it on its side and lifting straight up], but maybe lift up one end of it.)

I don't know if I could have gotten the head unstuck without either rotating the case in my hands and stopping its rotation against a hard surface, and/or heating up (or freezing) the drive with the hope of loosening the grip between the head and the platter. And if those methods wouldn't have worked then my current opinion is that the next best thing would have been to open up the drive and loosen it by hand (preferably just by spinning the platters) rather than attempting any other actions (such as hitting the side of the drive against a solid surface) - as anything else would have just created a force on the actuator arm that would have twisted the two tiny strips of metal that hold on to the head.

In the first image below you can see the two marks left behind by the parts of the head that contact the surface of the disk:

http://img148.imageshack.us/img148/9601/dsc04980yw5.th.jpg (http://img148.imageshack.us/my.php?image=dsc04980yw5.jpg) http://img223.imageshack.us/img223/5212/headbec3.th.jpg (http://img223.imageshack.us/my.php?image=headbec3.jpg)

Dwight Elvey
August 16th, 2008, 08:55 PM
All my successful attempts to overcome "stiction" have been by turning the drive spindle side up, putting a drop of 3-in-1 penatrating oil on the spindle shaft and letting it sit there for a while (hours).

Then hooking it up to a standard PSU and turning it on. in the majority of cases, this is enough to take care of the problem and letting it run in this position allows for full coating.

Sometimes, GENTLY moving the motor shell a couple of degrees, back and forth is necessary and several power on/off cycles.

After running for a few hours, turning it off, another drop of oil before it cools and another couple of hours of spinning.

I have drives that I did a couple of years ago and then set aside and about a month ago, I gathered them up, powered them and they still spin right up.

Hi
The problem you've had was not stiction, it was dry bearings. Any oil on the
surface of the disk would have destroyed the drive.
Dwight

Druid6900
August 16th, 2008, 09:14 PM
Yes, thank you, I know what the problem was and I believe that is how we are defining "stiction". The inability of the unit to spin up.

As for oil on the platter destroying the drive, I'm not too sure about that unless it was some highly corrosive substance.

I received a couple of ST-251s a while back, one intact and working and the other had no bubble top and had sat in a garage, open to the elements for over 20 years.

Also in the pile of stuff was a ST-225 with the infamous "blown custom IC" problem turning it into a paperweight.

I took the top from the 225, put it on the 251, screwed it down and powered it up. Despite the platters being tarnished from exposure, it formatted and took an operating system with no problem and has been running one of my test-jigs for over 6 months and still hasn't generated and flaws beyond the ones listed on the flaw map.

modem7
August 16th, 2008, 10:44 PM
Yes, thank you, I know what the problem was and I believe that is how we are defining "stiction". The inability of the unit to spin up.
In the context of hard disk drives, "stiction" refers specifically to the stiction that occurs the between read/write heads and the platters.

I guess it's like "screen burn". If I extinguished a cigarette on the face of my CRT, burning away some anti-glare coating, then put the CRT on eBay describing it as having some "screen burn", potential buyers have an expectation of what "screen burn" means. Well, I certainly did when I bought that cigarette damaged 21" CRT on eBay two years ago.

Druid6900
August 17th, 2008, 03:19 PM
Ok, if that's the definition, then I have never seen a case of "stiction".

modem7
August 17th, 2008, 08:11 PM
Ok, if that's the definition, then I have never seen a case of "stiction".
Neither have I.

I think that there have been cases where people determined that the spindle was not a spinin', and discarded the drive putting the problem down to "sticton", even though the problem may have had a different cause, e.g. failure of the driver electronics for the spindle motor.

Unknown_K
August 17th, 2008, 09:22 PM
I snagged an old Packard Bell P60 desktop earlier in the year when I tried to boot it up it made one hell of a rocket engine noise. I removed the HD and installed another one, and after I figured out how to kill the BIOS password it worked fine.

A few days after the machine was running I decided to rip the old HD apart (I keep the controller boards for parts and toss the aluminum case in the recycling bin) I noticed the head war ripped off and stuck to the top platter, the friction of the air as the platters spun up is what was making the noise. The bearings seemed to be fine since they moved freely. First time I ever had that problem.

QuantumII
August 17th, 2008, 10:57 PM
Just FYI, copy-paste from Webopedia and Wikipedia regarding HDD stiction:

Stiction: link (http://www.webopedia.com/TERM/S/stiction.html)
n.) Short for static friction, a condition in which a hard disk drive's read/write heads become stuck to the disk’s platters with enough strength to keep the platters from spinning, resulting in hard drive failure. When a computer is turned off, its hard drive's read/write heads park on the platter's landing zones. Under normal circumstances, the heads will lift off the platter when the computer's hard drive is activated and the platters rotate. Stiction typically occurs when a computer has been turned off for long periods of time.


Stiction and Hard Disk Drives link (http://en.wikipedia.org/wiki/Stiction#Stiction_and_Hard_Disk_Drives)
In the context of hard disk drives, stiction refers to the tendency of read/write heads to stick to the platters. Stiction is a complex and not very well understood phenomenon of hard disks. Stiction most likely occurs as a result of 2 properties of the platters, smoothness and magnetic forces. Once the heads have stuck to the platters, the disk will be prevented from spinning up and can cause physical damage to the media. Other forces considered as responsible for stiction include electrostatic forces and adhesion from the inherent stickiness of silicon.

In the early models of hard disk drives stiction was known to cause read/write heads to stick the platters of the hard drive due to the breakdown of lubricants used to coat the platters. During the late 1980s and early 1990s as the size of hard drive platters decreased from the older 8" and 5.25" sizes to 3.5" and smaller, manufacturers continued to use the same calendering processes and lubricants used on the older, larger drives. The much tighter space caused much higher internal operating temperatures in these newer smaller drives, often leading to an accelerated breakdown of the surface lubricants into their much stickier components. When the drive was powered off and would cool down(say at the end of the day when a user went home and shut off their PC), these now-broken-down lubricants would become quite viscous and sticky, sometimes causing the read/write heads to literally stick to the platter. The common solution to this problem was the counter-intuitive move of taking the affected drive out of the host system, striking it gently, but firmly on the side against a desk or something as laterally as possible and then re-install it in the host system. This would break the heads free of the goop long enough to power the system back on, have the drive spin up and recover whatever data could be retrieved off it. While the data was retrieved, the machine would be left on constantly so that the heat from the drive's internals would keep the decaying lubricants in a liquid state.

modem7
August 18th, 2008, 01:02 AM
The common solution to this problem was the counter-intuitive move of taking the affected drive out of the host system, striking it gently, but firmly on the side against a desk or something as laterally as possible and then re-install it in the host system.
I found the stiction "repair" section at http://www.microswamp.com/tech-notes/data-recovery/37-data-recovery/47-hard-drive-stiction-and-dta-recovery?345342a77c2a05fb40d72306ce9fe915=5a70b9ab 378ef6a4d2ac2737b7e36951 to be interesting.
The particular recovery technique given there will probably be the first technique I try if ever I encounter a case of suspected stiction.

Dwight Elvey
August 18th, 2008, 05:57 AM
Hi
I've seen these definitions of stiction. Although, there is a light amount
of lubricant on the surfaces, the actual problem as told to me by
a person that worked on the problem at Seagate was different.
What I was told was that as the densities of the drives increased,
the surfaces were becoming smoother and smoother. When two surface
are smooth and flat enough, they will actually bond together.
Anyone that has worked with machinest guage blocks knows what
this is. There are also articles about gecko lizards using this to allow
there feet to stick to surfaces like glass.
The problem was solved by making the surfaces rough enough
and still allowing enough data density.
The OP's problem was defiinitly some form of stiction. Enough
to rip the head from the mount.
Dwight

QuantumII
August 18th, 2008, 01:18 PM
It's settled then. The head stuck to the platter, and the disk is more or less beyond (economical) repair.

The best solution would be to get a drive off e-bay or similar :)

Anyone here interested in a working Zenith SuperSport/286 ? It does not have it's battery back or AC adapter anymore, but otherwise it's working fine :)

eupeptic
August 27th, 2008, 07:19 AM
For the curious, here are a few pictures of the platter and head taken using a 55 mm lens from an SLR camera taped onto the lens of the digital camera I was using (idea taken from AMagill on Flicker who took these detailed images (http://www.flickr.com/photos/amagill/sets/1819005/detail)):

http://img440.imageshack.us/img440/985/dsc05130gp9.th.jpg (http://img440.imageshack.us/my.php?image=dsc05130gp9.jpg) http://img514.imageshack.us/img514/4155/dsc05133om7.th.jpg (http://img514.imageshack.us/my.php?image=dsc05133om7.jpg)

The first is focused on the mark closer to the spindle and the second is focused on the further mark. Note the buildup of some kind of fluid (which I'd assume is lubricant based on the explanations I've heard for stiction) around the edges of where the head was in contact with the platter.

http://img440.imageshack.us/img440/5121/dsc05126aa2.th.jpg (http://img440.imageshack.us/my.php?image=dsc05126aa2.jpg) http://img514.imageshack.us/img514/5849/dsc05132bg8.th.jpg (http://img514.imageshack.us/my.php?image=dsc05132bg8.jpg)

In addition to the spot where the head stuck to the platter, there are also these two separate locations where some of that fluid was deposited but the head didn't get stuck. I'd assume that these occurred on the occasions when I powered up the laptop for just a few minutes to see if it was working. (I'm not sure why these marks are shaped differently than the head...)

http://img82.imageshack.us/img82/640/dsc05150wm6.th.jpg (http://img82.imageshack.us/my.php?image=dsc05150wm6.jpg)

A close-up of the head. Note there's more fluid visible on the trailing edge of the head (which is closest to us).

http://img82.imageshack.us/img82/496/dsc05189da3.th.jpg (http://img82.imageshack.us/my.php?image=dsc05189da3.jpg)

Just for kicks, a comparison between this head and the head from an IBM Travelstar 3XP (http://www.hitachigst.com/hdd/support/dlga/dlgades.htm).

http://img186.imageshack.us/img186/9025/dsc05083mv8.th.jpg (http://img186.imageshack.us/my.php?image=dsc05083mv8.jpg)

And the side of the drive that we're looking at here faces up when it's installed in the laptop. (The drive slides into a case that covers everything but the part on the right.)

QuantumII
September 1st, 2008, 05:10 AM
Nice pictures :-) Did you get a replacement drive yet ?

aaaall
September 1st, 2008, 11:58 AM
On the old hard drive that was sitting there for a while, remove the controller board, slightly move the spindle by hand to loose the heads that were sticking on the platters before fire it up. If you plug the power in, turn it on (without loosing the heads) Heads will break off. If it broke, scrap the hard drive, replacing head is not a easy job. NO dust, oil can be in the hard drive. I am not YET talking about re-write the servo track, align the head assy.

MikeS
September 1st, 2008, 03:35 PM
[QUOTE=eupeptic;69911]For the curious, here are a few pictures of the platter and head taken using a 55 mm lens from an SLR camera taped onto the lens of the digital camera I was using (idea taken from AMagill on Flicker who took [url=http://www.flickr.com/photos/amagill/sets/1819005/detail]these detailed images
QUOTE]
----

Interesting idea; I'd love to read/see some more details on how to tape that SLR lens on to the digicam - tell me more...

m

eupeptic
September 3rd, 2008, 02:53 PM
QuantumII, I did try running an IDE drive in it and it boots v e r y . . . s l o w l y . . . (I probably only waited 30 seconds or so as some of the characters were only coming on the screen one at a time.)

MikeS, the camera (a Sony Mavica MVC-CD350 (http://www.dpreview.com/articles/pma2003/Images/Sony/IMG_7894.jpg)) was configured as you see below with an old Mamiya / Sekor lens (there is a UV filter on the SLR camera lens that I probably could have taken off to so the edges of the pictures wouldn't be cut off as much, but I didn't think of doing that as I was happy with the results), and I had it on a tripod in a room underneath a fluorescent fixture (one that uses the straight lamps). Objects are in focus about 1 inch from the end of the lens (I often use that lens as a magnifying glass as it gives a very clear and highly magnified image, but I had never thought of sticking it on a camera as I've tried taking pictures/video through a microscope and binoculars without achieving any good results), and the camera seems to be able to focus by as much as 5 mm or so when it's zoomed in all the way (if I don't zoom in the image is in a circle surrounded by black from the 55 mm lens). While taking pictures I was using the timer as the exposure was often down in the 1/4 second range (at ISO 100), and a few required a 1 second exposure (I wasn't using any flash). Getting the camera to focus where I wanted wasn't that easy (notice that the first image of the head is not centered on the head, but on the part I wanted in focus), but it worked fairly well most of the time.

http://img98.imageshack.us/img98/2156/pic0030vx9.th.jpg (http://img98.imageshack.us/my.php?image=pic0030vx9.jpg) http://img231.imageshack.us/img231/1926/pic0027kx5.th.jpg (http://img231.imageshack.us/my.php?image=pic0027kx5.jpg)

MikeS
September 4th, 2008, 07:57 AM
Thanks for the info; gonna check it out.

m

MikeS
September 4th, 2008, 08:38 AM
By the way, I have a Conner CP-3000 here and should also have a few WD AC-280s somewhere, both of which are rumoured to be compatible with at least some SS-286 models; if you're interested I could check 'em out to see if they still work.

What make/model IDE drive are you (slowly) running now?

m