Will freezing your hard drive help recover its data?

Question

There seems to be (anecdotal) evidence that putting your hard drive in the freezer can make it readable again, at least for a short time:

I guess that by doing this, some metal parts in the HD could contract, putting back in place defective parts, and making everything work again for a few minutes. [Source]

But data recovery expert Scott Holewinski advises against it in this video (YouTube), saying

... putting your hard drive into the freezer is the WORST thing you could possibly do in your attempt to recover your data as it can actually render your HD completely unrecoverable (due to water condensation)

But people who say it does work usually recommend wrapping the HD in a sealed bag to avoid condensation (see picture above).

My question:
Is the method of freezing your HD for data recovery viable?
Will sealing the HD in a ziploc bag actually help avoid the condensation problem?
(AFAIK there is air inside the HD)

Haven't really looked at this, but I'm predicting "No way." I think this is probably a spin off from a method to recover data from *RAM*, not standard, platter-spinning HDs. It's a well documented effect that allows a "cold boot attack" -- [LINK](http://citp.princeton.edu/memory/). Perhaps these folks are extending that toward other types of storage? — Hendy, Jun 25 '11 at 13:47
I'll add to the anecdotal evidence. I had a hard drive go south on me to the point that it would no longer boot. I had nothing to lose so I tried this approach and it worked. At least it worked long enough for me to backup some important files off of it. I think it would depend highly on the reason that the hard drive failed, but based on my own experience I'd call it plausible. — JohnFx, Jun 25 '11 at 16:17
"Is the method of freezing your HD for data recovery viable?" - There's anecdotal evidence that "it worked for them": which implies that the answer is, "sometimes". My guess as to how that's possible is that if the head is touching the platter, freezing it might cool the air inside enough to make the air denser, which could allow the head to 'fly' a little better. — ChrisW, Jun 25 '11 at 17:12
Sadly I only have anecdotal evidence, but we used this successfully 8 or 9 times in the mid nineties When a hard drive had died, this did sort it out long enough to retrieve the data off it, before it was then discarded. We also had 4 attempts that failed. I don't have any evidence from more recent drives. We did also successfully use the de-stick technique of smacking a drive off a concrete block a few times as well, but don't try it at home:-) — Rory Alsop, Jun 25 '11 at 17:22
Belongs on Superuser, exact duplicate of http://superuser.com/questions/1078/harddrive-in-the-freezer-ever-work-for-you ;) — Andrew Grimm, Jun 26 '11 at 02:40
I'd suspect confirmation bias - in other words "put it in the freezer -> it didn't start working -> never mind, didn't really expect it to" might be a much more common outcome than "put it in the freezer -> it did start working -> OMG IT WORKS", but the if latter occurs, you are *much* more likely to remember it. (I'm not aware of a study on this specific recovery technique, though) — Piskvor left the building, Jun 26 '11 at 16:59
Complete opposite anecdote that adds nothing but is interesting: I've had a graphics card burn out on me four times now, and each time it's happened I've baked it in the oven and revived it. 380 degrees F for 10 minutes liquifies the solder enough to reseal any cracked joints. Just make sure you follow a guide and take the card apart; you don't want to bake the wrong parts. — erekalper, Jun 27 '11 at 14:40
@Andrew Grimm: nice find... but their answers are super disappointing. Just anecdotes. I *would* like to find evidence supporting the answer that it has to do with one type of failure -- bearings stuck that prevent spinning up. *That* would answer the question by showing the mechanism. — Hendy, Jun 27 '11 at 16:48
@Piskvor: Do you mean [Publication bias](http://en.wikipedia.org/wiki/Publication_bias) rather than confirmation bias? — Andrew Grimm, Jun 27 '11 at 23:00
@Andrew Grimm: Good catch, I do; wonder what I was thinking :) — Piskvor left the building, Jun 28 '11 at 07:54
Wrapping the hdd with sealed bag doesn't help with condensation. The condensation happens after the hdd was taken out of the fridge. I believe the condensation happened inside and outside of the hdd. — lamwaiman1988, Jun 29 '11 at 02:28
Ice forming inside a HD must be a bad thing, as tolerances are tight inside there and the thing spins at ~100 revs/sec. — Paul, Jun 08 '13 at 02:55

Chad · Answer 1 · 2011-06-30T13:17:08.767

24

I speak as a professional who has worked as a Data Recovery specialist for a firm that specialised in this.

There are three common types of hard drive failure:

Control Board failure: This is where there is a short or surge that damages the controller board of your hard drive. The symptoms of this are the drive not being found by the BIOS, the drive not "spinning up", sometimes an odor of ozone (if the short is active). Freezing will not fix this failure. Replacing the board usually allows for full recovery of the data. The difficulty is finding a board that works with your exact model of hard drive. If you have one, it's very simple to replace.

Motor Seized: This is the motor that causes the hard drive to spin. It was fairly common in the early hard drives. Parts overheat and expand causing the parts to wedge themselves in a position that doesn't allow the drive to spin freely. Often you could put these in the freezer and it would cause those parts to contract and release allowing the drive to work again. The miniaturization of the motors have reduced the frequency of this failure, and made it so that the damage from them is often irreversible. The data is still on the platter and can be recovered. Symptoms: Drive is recognized by the BIOS But does not spin up. Sometimes you can hear the motor trying to move but it is a constant whine (older drives only). Data recovery places do not want you freezing these for several reasons:

it is an easy fix, if you know how to fix it and have the time, so they can charge a lot of money and recover all of your data (win win),
if it is not a motor failure and is other mechanical failure you can cause data corruption(see below)

Head Crash: This is the noticable kerthunk kerthunk kerthunk that anyone who has ever heard it remembers. The drive seems to spin up, thunks 3 or 4 times then spins down - rinse & repeat. Sometimes it will even boot up. This thunk is the drive head making contact with the platter and being propelled away (momentum) every time the head touches the platter it scrambles some of the data on it. The more often the head crashes the more data is destroyed, the less likely to recover what you want. No amount of freezing is going to help here. There is physical damage to the parts that are responsible for reading the data. In order to recover data from this we have to replace the drive heads in a positive flowbench (a poor-man's clean-room) and hope the disk damage does not cause them to crash again. These are generally a bit more interchangable.

Evidence: http://www.dataclinic.co.uk/hard-disk-failures.htm

edited Jun 30 '11 at 13:17

answered Jun 27 '11 at 17:46

Chad

9,099
6
49
96

5

Unfortunately, anecdotal evidence has been deemed harmful to this site. [Back up your answer with sources or remove it](http://meta.skeptics.stackexchange.com/q/5). – Sklivvz Jun 27 '11 at 18:06
3

Nothing on your reference page makes reference to freezers. – DJClayworth Jun 27 '11 at 19:20
Updated the post with references. – Chad Jun 28 '11 at 12:46
Does he need actual evidence of the freezer contracting metal to allow the parts to move freely? I feel like that is common sense, he referenced the part about heat expanding the metal, thus freezing may make it work by contraction. I mean how far does the reference thing go? Should he stick some metal in the freezer and tape it contracting... sometimes common sense needs to take a front seat when the other references logically lead you to a conclusion... – Supercereal Jun 28 '11 at 14:01
2

No reference is needed to show that freezing metal causes it to contract. A reference is certainly needed to show that this has a noticeable effect on making broken hard drives work. – DJClayworth Jun 28 '11 at 14:25
@Chad I still see nothing in the references about freezing. Neither the "Motor seized" link nor the "Evidence" link contain any word starting with "freez". Please tell us where we can find actual stuff that talks about freezers. – DJClayworth Jun 28 '11 at 14:30
I explained how it works, why it works, and when it can work. I dont know what more I can do. I do not have a reference to a page that says it works because it doesnt work on modern drives. I also explained why that is. – Chad Jun 28 '11 at 15:47
If the fact that freezing helps isn't confirmed by some official evidence, how can you say, as a skeptic, that this procedure works? Your answer has certainly improved, however it is still far from being acceptable here. – Sklivvz Jun 28 '11 at 17:52
@sklivvz - Because I explained the process where it is effective- when parts are overheating and seizing up. I explained why it can work then. And I explained when it does not and will not work. – Chad Jun 28 '11 at 18:13
2

Chad, all you have provided is a *theory* on why it could work. This is not a sufficient condition for it to be actually happening. I can provide a slew of competing theories that equally "prove" the opposite: "freezing contracts the metal parts of the drive outside the operational parameters and damages them further"; "Internal condensation while thawing fries the electronics"; etc. – Sklivvz Jun 28 '11 at 18:18
@Sklivvz - That happens too. Just because there are failure conditions does not mean that there are no success conditions. It **it CAN work if overheating has caused the parts to expand and seize in a position that when the heat returns to normal levels the parts are still seized** If recovering your data is more important than saving money send it to a professional. If you would like to get the data off and you have an old drive that is WHINING but you dont mind losing it sure go for it. – Chad Jun 28 '11 at 18:24
4

You need to prove it with references. Fix your answer or remove it. I am not wasting any more time repeating myself. – Sklivvz Jun 28 '11 at 18:30
@Chad let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/649/discussion-between-sklivvz-and-chad) – Sklivvz Jun 28 '11 at 18:31
for the second fix you can solve it without freezing http://www.sentex.net/~mwandel/tech/repair.html – ratchet freak Jun 28 '11 at 22:17
Please do not remove the banner unless you add references. Someone else might want to help since you don't want to fix the problems with the post. – Sklivvz Jun 30 '11 at 13:13
Anecdote vs anecdote -- At work we've managed to recover data off a few "kerthunk kerthunk kerthunk" drives by freezing for an hour or so. Their was still data corruption, lending support to the "crashing head destroys data" part of your statement, but I don't think you can state with authority that no amount of freezing will help. Overall, unless someone digs up a study that specifically looked into this, I doubt we'll have anything but anecdotes for this claim. – Saiboogu Jun 30 '11 at 14:33
@saiboogu - The only thing i can think of is those had some sort of siesure that caused the heads to crash. So the freezing allowed it to break from the seisure. Had the head crashed on the MBR or file allocation tables (Happens alot) booting up again would have done nothing more than cause more file corruption. – Chad Jun 30 '11 at 14:43
2

@Sklivvs you want evidence that proves a negative. Because freezing is not a guarenteed fix failure to work for one does not disprove. Because there is also a chance that something else caused the recovery. All I can do is explain when it can work with logic. You can not prove it because taking apart the drive to identify the problem changes the problems. I have seen the evidence of seisure on motors and bearings. I have seen drives with siezed motors return to spinning and we analysed why. – Chad Jun 30 '11 at 14:48
1

Maybe there should be a vote from experience, because I'm no professional in HDD recovery, but I did recover data from a failed Maxtor a long time ago, by freezing it. It worked just long enough to read some important data from it. I also consider Chad's answer should have been the accepted one. Also, how is a datarecovery.com reference better than a free professional's word? datarecovery.com is trying to sell stuff, so they will not help anyone recover for free (freezing a drive). A free professional will. – oxygen Jun 26 '12 at 09:16
1

My experience with this is it's something to try right before you write the drive off, but treat it as a crude desperate measure that often does more harm than good. Yes, it can sometimes work with the right sort of mechanical problem, but in the situations where it works, a data recovery company has nearly 100% chance to get all of your data back, whereas freezing the drive will destroy it for about a 50/50 shot at getting a couple minutes of life out of it. If you try it, consider it a last resort and work fast. – Stephanie Jul 28 '12 at 20:03
http://www.hddrecovery.com.au/PDF/200ways.pdf says about cooling old drives – woliveirajr Sep 26 '13 at 13:01
The links you provide are merely definitions and don't support your answer (and two of them are broken now). Can you fix this? – Sklivvz Feb 11 '14 at 10:19

score 14 · Accepted Answer · answered Jun 30 '11 at 21:20

TL;DR: No, it's not a viable method, it will likely cause damage and will not likely work at all.

Here's what the largest data recovery firms say on the matter:

datarecovery.com

Stamped on pretty much every page of the site:

Please keep in mind that hard drives are delicate instruments. Like any other delicate instrument, if they are not handled with care or properly cooled, you may experience a failure. Hard drives are sensitive to heat, rapid temperature changes, static electricity, power surges and physical shock. We estimate that most hard drive failures occur due to operating conditions, not manufacture defects.

Which pretty much means, don't mess around with temperatures because HDs are very sensitive to that.

In fact, according to krollontrack.com, hard drives are so temperature dependent that thermal expansion (and I am adding: thermal contraction) are relevant to the design of the base casting assembly:

Today’s hard drives have no room for errors when it comes to platter and head alignment. The tolerances are so exacting that hard drive manufacturers even design ways to keep the Base-Casting Assembly, where all the components are attached to, from shifting due to high temperature situations. For instance, one hard drive manufacturer of high performance SCSI based drives actually designs their Base-Casting Assembly with pre-stress points. The assembly does not line up from corner to diagonal corner—it’s pre-torqued. When the casting assembly heats up, the unit actually twists back (thermal expansion) into a true line-up from corner to corner.

Unsurprisingly, the correct advice abounds on data recovery sites, for example datarecoverylabs.com:

If a drive is making unusual mechanical noises, turn it off immediately and contact a data recovery specialist.

securedatarecovery.com—boasting high-profile customers, from Nasa to Microsoft—also gives the same advice:

Stopping all use of the damaged equipment or medium is the single most important step in any data loss scenario.
Attempting to perform data recovery on your own is very dangerous - if you are not absolutely sure of what you are doing, you might jeopardize the safety of your data. Do not attempt to use data recovery software, and never try to disassemble the hard drive on your own. Continuing to use your hard drive while it is in a degraded state can endanger your data and make the data recovery process more difficult.

But I think nobody makes it more explicit than datarecovery.net:

DO NOT put your drive in the freezer and then try to spin it up. It is possible that moisture has condensed on the media surfaces. This WILL cause head contact if it has and will destroy the drive.

DO NOT listen to your friends or continue to look for home remedies on the net such as the one mentioned above, seek professional help if you value the lost data.

As a side note: not a single one boasts "cryogenic units" or anything similar. All the most reputable ones boast clean rooms and privacy/security standards.

+1 Persuasive. I misread "viable" in the OP as meaning "possible" or "having a mechanism", rather than as "probable" or "profitable". — ChrisW, Jul 01 '11 at 03:44
So, companies that want you to spend money on their services are telling you not to try to fix it on your own? I have a bit of a hard time trusting these sources. — mmr, Dec 14 '11 at 15:47
@mmr: agreed. However. much more reliable than the anectodal evidence presented in other answers. — Sklivvz, Dec 14 '11 at 16:22
Data recovery companies are not a good reference, especially when they spread fear so you buy their services, instead of fixing it yourself. I freezed a drive, and recovered data from it. Am I a sinner? — oxygen, Jun 26 '12 at 09:21
I agree with @Tiberiu-IonuțStan that data recovery firms aren't what I would consider and unbiased resource. However, something else I would be curious about is the age of the hard drive itself. Hard drives 10 years ago where not built the same as modern hard drives and what is try now may not have been true then. — rjzii, Jul 09 '12 at 01:41
I disagree the companies are not a good reference. While they may have some interest in advising people not to try and fix their drives, their services are very expensive and targeted to valuable data and companies. I think they are more concerned with having simpler cases to handle... Put it this way: when doctors say "don't move in case of injury", does one quip "doctors say this because of self interest?" @rob — Sklivvz, Sep 04 '12 at 07:59
@Sklivvz - This might be a better conversation for meta but trying to compare the advise of a doctor versus the advice of a company is kind of like compare apples to bananas as doctor's are supposed to give unbiased advise where as the company is under no obligation to be truthful. Don't forget it wasn't that long ago that we were told that smoking is good for you. — rjzii, Sep 04 '12 at 11:52
I agree with Skliwz. If it was reliable someone would be making money with it. It is unlikely that there is a conspiracy by all these companies to hold fixing drives by freezing down. There is an xkcd comic for that one (technologies that if they worked would make good money), but cant link it from work. — Stefan, Sep 04 '12 at 12:31
If freezing drives increased the likelihood of successful data recovery (which, IIRR, is a low likelihood to begin with), would not the companies recommend you do it, and have you ship the drive to them frozen? — Mathieu K., Feb 15 '16 at 21:50

Will freezing your hard drive help recover its data?

2 Answers2

Linked