The idea that freezing an unreadable hard disk could bring it back to life has been around for a long time. Ordinarily I’d say “No way, don’t do it”. If you’ve lost your hard disk the last thing you want to do is mess with it in any way. Take it to an expert; a real expert that is – not some shop on the high street with a sign up saying “Data recovery and virus removal”.
However, this story doesn’t go away. And there is a grain of truth in it. I remember some mid-1990’s drives did benefit from the freezer treatment – it shrunk the platters slightly and realigned the bearings. Or that was the theory. Anyway, it sometimes worked. Back then. However, luser forums are full of stories where people have used this technique on modern drives and claimed success.
On my desk at the moment I have a 160Gb SATA Maxtor, vintage 2002. And it wouldn’t read. This is partly because parts of the platter are now unreadable, and partly because the NTFS filing system is mangled. I know this kind of stuff. After a couple of days I reckon I’ve got all the blocks off it that it’s going to yield, but repeatedly retrying the bad sectors. Some read eventually, other’s don’t.
Now it’s an observable fact with dodgy platters on winchesters that they’ll sometimes, briefly, come good. It’s worth powering down the drive, letting it cool and trying again later. It’s even worth trying it at different angles. Having a few fans around it to keep it cool is a good idea anyway. Using these techniques I’ve recovered about 80% of the bad sectors, with about 2500 left that aren’t doing it. A good candidate for the freezer perhaps? Well, as a last resort – I’ve tried the drive at room temperature, working temperature (about 35C) – where’s the harm in trying it sub-zero?
I mus emphasise here – this is a last resort – 99.999% of the data is off it and it’s stuck a that. There’s nothing to lose.
So, in to an airtight box and into the freezer with it.
Did it work? No! In fact the whole drive ended up unreadable. I’d put that down to the condensation. If you want to knacker a drive without leaving marks, condensation on the platters is a good way to go.
Leaving it to get back to room temperature did result in it coming back to life, although I only got two more sectors back.
So why do all these people keep insisting it worked for them? Coincidence, I reckon. If they’d left the drive alone for 12 hours it might have started working anyway. They often do, even for a brief period. This drive had proved that some sectors were heat-sensitive; more could be read while it was still cool and the recovery rate dropped off when it was warm. If it didn’t help with this drive, I seriously doubt it would work on other candidates.