sorry wrong topic.I have finished over 2000 drives until today
few I even had to freeze.
Just few weeks ago I was working an 1988 Fujisty for Cape Town airport radrar system.120Mb toshiba I had to push it up down in order to work.
But there is only one drive that has firmware problem that can have some succes with freezeing that is maxtor 540-4k,so let thet freeze they will not lean a thing.
I had a blown motor driver should i freeze it?
Hi All,
Just been reading this topic with interest.
I work for a data recovery company in the UK, I am forever hearing of success stories using these methods, and in all honesty I can say in the early days these methods were used with a very limited degree of success on certain drives.
The point I must make is that the results of our testing on temperature tolerances on drives suggests that freezing the drives should make no difference, and vapour issues 'can' (rarely) be detrimental to the hard disk.
I agree with all that has been said, if you have nothing to lose, give it a try, however a note of caution Modern drives using FDB (Fluid Dynamic Bearings) have during testing produced a high number of seizures due to lowering the temperature below manufacturer specification. A seized motor is a serious problem, and one not easily rectified - so be careful.
Chris
notanotherbob nailed it.
Usually with older drives and not something I've seen lately, the tracks on the platters were tracked by the physical distance from the hub. If the platter changed size the head couldn't find the track. Freezing the drive caused the platter to shrink a little and thus could be read until the drive heated back up. You usually had just enough time to copy everything before the drive went unreadable again. Sometimes the platters were actually smaller than usual - maybe the drive was formatted after it was already running for a while in a cramped server and now it is cold - and the platters need to heat up to be the right size to track correctly. Nowadays, the tracks aren't tracked by location but rather by watching the signal on the platters. The head will adjust to find the track, so even if the platter swelled the head would still track correctly. Ceramic platters aren't as susceptible to temperatures as the older metal platters, also.
The worse case I ever saw was a drive that warped due to overheating. We built a jig and applied pressure on two opposite corners of the drive's case until it started reading again.
The clicking noise so common in drives is actually the head losing sync so it can’t see the track anymore. It goes back to the center of the drive (click) and then back to the track it was trying to read (click), and keeps repeating this. The clicking is really the head banging from side to side, but since it is small it sounds like a click. With current drive technologies, this is caused by a defect in a track that makes it unreadable. The above tricks don’t really apply, because the sync isn’t being lost due to platter expansion/contraction but rather due to a defect in the platter itself.
As for "physical manipulation", this was usually used to solve stiction - meaning the head was stuck to the platter not in the place it was supposed to touch the platter. Sometimes just a real quick, j**k-like spin of the drive was enough to add a little extra inertia to get the platter to spin up. Sometimes a 'tap' to the drive case (as you described as 'percussive maintenance') would bounce the head up off the platter. Sometimes the best bet was to remove the top cover and spin the platters by hand (touch the hub - never the platters – rotate so that the platter passes under the arm and then the head, so the head is pulled not pushed – never rotate the platters the other way – wear anti-static wrist wrap – wear latex gloves so skin oils don’t get on platter; anti-static if you can get them – work in the cleanest environment you can find – pray a lot). If this worked the drive would work long enough to get at least some of the data transferred off and then the drive is trashed due to the introduction of dust. If this didn't work (applies to all of these techniques actually) the most likely result is either the head will never 'unstuck' itself and thus the platters will never spin, or the head will rip off the arm, the arm will gouge the platters,8 and life will get a whole lot worse.
Whatever the outcome, these measures are usually used when all the polite measures have been tried and the drive is on the verge of being considered a brick anyway - so what do you have to lose? With all of these techniques, the drive should be trashed once the data is off anyway. These are all temporary fixes to read the data, not repairs that make the broke drive usable again.
Regarding you question about appropriateness of this topic to forensics
Computer Forensics = Data Recovery + lawyers
Hi All,
I agree with all that has been said, if you have nothing to lose, give it a try, however a note of caution Modern drives using FDB (Fluid Dynamic Bearings) have during testing produced a high number of seizures due to lowering the temperature below manufacturer specification. A seized motor is a serious problem, and one not easily rectified - so be careful.
A very good point - You don't really want to 'freeze' the drive, you just want to lower the temperature a bit.
Simmo, can you recover from siezed motor bearings on multiple platter drives ??? i have got a three platter drive with a siezed motor….can i do this do you think ??
Thanks minty.
OK I know it's anecdotal, but over the years I have had great success in getting images off difficult drives by reducing their temperature. One situation springs to mind
I was on a client site in Madrid attempting to image a drive which contained evidence. I was on my third attempt when I became aware that the disk was running unusually hot (the weather wasn’t hot, in fact it rained for the 3 days I was there, and I had two broken ribs – but that’s another story!). I placed the drive over one of the air-conditioning vents around the edge of the room and turned the thermostat down. The client was a little confused to come back in and find myself and my colleague sitting huddled up with our coats on, but he was mighty impressed when it actually worked and we got a good image to work from followed by a full confession from the suspect employee!
I have also had success imaging difficult drives by standing them on their edge rather than lying flat. Maybe it shouldn’t work… but it has!
The client was a little confused to come back in and find myself and my colleague sitting huddled up with our coats on
Hmm, this conjures a picture in my mind almost as if you were practising Voodoo!
or re-compiling the latest Linux kernel.
😉
I've also had success with cooling the drives off. And sometimes the position of the drive is what makes it work. Both of these are exactly why I make 2 images of everything before shutting the original drive off.
Although I dont recommend the freezer trick, as it can cause 'stuck' heads to shrink and therefore scratch the platters, we at Disklabs have a freezer to assist us in data recovery.
Generally, its just used for storing cold drinks, but occasionally, we get a drive that gets so hot during the imaging process, that it falls over, or causes internal problems. At this point, we generally use the fridge overnight to cool the problem hard drive down and then attempt the same proceedure again, (knowing that we are likely to get a bit more life from the hard drive than previously). Please note that we only use this for drives that overheat.
This issue with dropping a hard drive is dodgy to say the least! When the hard drives head stick it is known as stiction. In the old days IBM drives were reknown for this, but it doesnt happen as much as it used too.
I hope the above was a bit of use to you.
Regards,
I think what you guys are referring to is the old drives which contained proximity sensors for speed control.
I did this many years ago where the fault was with the proximity sensor dislodging. Due to this the disk read/write timing altered just enough for the disk to fail read attempts unless it was warmed up - the same goes if the sensor had moved the opposite direction - it would have needed to be cooled.
Of course, these days with such large disks and so much data per inch large temperature swings can cause metals to enlarge or shrink potentially causing issues with recovery.
This does pose a question about forensic imaging though; at what point was the data 'correct' when cold or when warm?