HDD Resurrection Using Freezer | Hacker News Reader

HDD Resurrection Using Freezer

krakonos.org

17 pointswslh13 comments

I've always been exceptionally skeptical of the freezer claim due to a lack of any real comprehensive explanation as to how it works or any actual study that gets even a base number of failing drives that were temporarily recoverable as a result of the freezing. I'm not convinced that the result would be any different if the drive was simply allowed to rest and tried later.

Every explanation has talked about parts expanding and contracting, but given how easy it would be to demonstrate this by simply disassembling the drive and doing before/afters, I feel incredibly skeptical of such a claim and see it more as a wild guess than anything.

All that being said, even though I see it as voodoo, freezer is still my last stop with a drive when everything else fails, though I still let the owner know that I'm not expecting any results.

That's what I was gonna ask -- what is the logic by which this method works? I understand how "baking" a graphics card or other piece of hardware can extend its life, but not how this saves HDDs.

From what I understood, the "freezer method" could work only in case of mechanical problems caused by wear. For example physical wear of the various hard drive bearings could lead to miss alignment of the read/write head and the tracks on the disk, therefore forcing the drive to operate at a lower temperature than "normal" tightens these bearings and reduces the miss alignment at least until the hard drive heats up again.

Same it's kind of an old school myth. It gives you something to do I guess.

That said, it worked for me once, I was able to mount my old spinning disc for a few minutes using ice packs attached to both sides of the drive and tied with elastics :)

It's not a "myth" I've done it to 3 drives when their heads got stuck 2 of those it worked afterwards one of those worked for over a year.

My personal theory is that it shrinks enough to release the seized heads and unlike striking on the side of the drive with a rubber mallet you don't run the risk of damage to the plates due to a head strike. It could also be that the shrinking fixes/temporarily aligns something in the surface mounted components, maybe it breaks some tin whiskers that grew between the solder balls under one of the chips or w/e but I've seen electronics revived before after being frozen and brought back to temperature usually that doesn't last that long after.

I've brought back to life quite a few HDD's via various means the most common of which was replacing the controller board with an identical one, mechanical failures were harder to fix but the freezing thing does seem to work in some cases.

I don't doubt your technical prowess for repairing HDDs, but my skepticism is based largely in that despite how overwhelmingly well known it is, it's never been really tested in a formal setting and the explanations have never gone beyond just "expanding and contracting", with no deeper explanation. Aside from everyone's stories about how it worked, there really isn't a lot of good solid knowledge about whether or not it actually does anything.

I'm not doubting stories where it went into the freezer and worked; quite the contrary, I'm more than willing to believe that was part of what was done to do it.

What I question is whether or not it was the actual freezing process that did anything, if there actually is a physical change that occurs beyond people's speculation.

It would seem to me that if there is a physical change going on with the drive, it should be visible, or at least a difference ought be visible in a before an after, save a few minor parts not visible to the naked eye or without further disassembly.

The fact that there really hasn't been any research on this, even amateur, despite the number of server farms, kind of makes me wonder if it's not just a bit of old voodoo.

Freezing (as well as sharp acceleration) can unstick the mechanical parts of the HDD quite often the arm that moves the heads up and down the platter. In my case i could hear the "ticks" of the arm stepper motor they were gone after a day in the freezer. And I've seen an iphone that would not work you put it into a freezer for 1-2 hours, take it it turns on, after about 1-2minutes it dies again and you could repeat it. There are probably so many things that temperature could affect when it comes to electronics that I'm not sure if there is even one thing that a research could easily focus on.

Well, that's sort of my point - with how many different items "could" be studied and quantified, not a single study has been produced to measure even one aspect just to say "we noticed no discernible change after the freezing process". There isn't even really a good baseline statistic for what the success rate is.

In the 30 some years I've been hearing the freezing idea, I find it really peculiar that nothing besides anecdotal evidence, not even amateur studies of just like a server farm testing failing drives or an IT desk recording their attempts that gives a sample size of more than ~10ish.

Conjecture and theories are great, and it's good to hear that these drives came back online. But, without exaggeration, I've tried this on hundreds of drives over the course of my career when doing IT support, and I can think of maybe 2 instances when after the freeze, the drive spun up long enough for data t be recovered. And everyone's stories seem to be this way - just this one or two times, after a freeze, it worked. There are no stories where a company comes out and says "man, freezing drives has saved our butts and customer data tons of times!"

I can't stress enough that this isn't a declaration that freezing doesn't do anything. But I feel the skepticism is absolutely appropriate given the voodoo qualities of the process and the overwhelming failure rate the process has. The lack of research on it and the fact that it's just a "hail mary" for most people rather than a cemented procedure just seems to me like the procedure is just ceremony, and that if put to any sort of actual trial, the success rate would fall well within the range of chance.

Why freezing can work sometimes but generally won't: http://www.datarecoverygroup.com/articles/freezing-your-hard...

Anecdotally, I've tried this with multiple drives previously and it's worked several times.

The caveat being that it'll only work for a short time (basically just long enough for you to extract data). I've also had to refreeze a drive several times to get things out of said drive.

It's best to start replacing drives as soon as SMART shows an error. Got a drive coming next week just for that reason.

Anecdote here. I had an ~8 years old 500GB Western Digital hard disk which spent the last 3-4 years on a shelf and when I tried to reuse it, after 2-3 uneventful months, it started to "refuse" to spin up at boot time and, after several seconds of retrials, BIOS would print the classic "boot failure" message. Sometimes it would spin up after several reboots but after a while it refused to spin up at all so, after a quick google, I got to the "freezer" method and to my surprise it worked. After being left in the freezer for ~1 hour (wrapped in plastic) it spun up instantly and functioned normally, long enough to make a backup of some not very valuable files from it (after ~8 years, I was expecting failures). So "freezer method" could work as a measure of last resort.

After trying 4-5 drives with this trick, only 1 was able to recover data using this technique.

I have tried this trick a couple of times without success.

Reply on news.ycombinator.com