A hard drive that has crashed or cannot mount can be returned to life by leaving it in the freezer overnight.
Another "digital urban legend?" I've read versions of this bizarre story regarding freezing a crashed hard drive in blogs, in Usenet group threads and on individual websites. The details vary, but the basic idea is the same: If you have a hard drive that has errors, won't boot or doesn't even spin up, you may get it to work by putting it in your freezer for a day, then firing it back up again.
Not every hard drive having problems responds to this freezer treatment, but in some cases it does seem to work.
But why would putting a crashed hard drive in a freezer revive it? I figured the people who might know the answer to this question would be the folks who perform data recovery for a living. So I contacted Jim Reinert, vice president of data recovery and software products at Kroll Ontrack. According to Reinert, freezing a faulty drive might have been beneficial once upon a time, but is no longer recommended.
Why freezing a hard drive might help
 |
|
|
|
|
"When the drive is put in the freezer, the materials shrink back to their normal orientation and the connections are reliable."
Jim Reinert
VP of data recovery and software products, Kroll Ontrack
|
|
|
|
|
|
|
|
|
|
I'll list three technical reasons that freezing a malfunctioning hard drive might fix it, along with Reinert's thoughts about why this practice might not be a good idea any more.
1. Board expansion due to heat. "The chips on a [printed] circuit board (PCB) and the motor in the hard drive itself generate heat, and most materials expand slightly when heated," said Reinert. "In some cases, this can cause the solder connections on the circuit board to be unreliable. When the drive is put in the freezer, the materials shrink back to their normal orientation and the connections are reliable (until the circuit board heats up again)."
Electronics failure is the most likely symptom to be addressed by freezing a drive, since chilling the electronics can delay thermal expansion of weak solder joints. That said, modern drives use far fewer components on the PCB than before, which means less heat and fewer individual points of failure. Reinert also noted that nearly all modern drives use fluid-bearing motors which produce less friction, and therefore less heat than ball-bearing motors."
A drive that's been around for several years might still be using a ball-bearing motor, but most drives sold in the past few years use a fluid-bearing system. A white paper from Hitachi discusses the benefits fluid-bearing motors bring to hard drive designs.
2. Alignment of the head stack. The head stack is the assembly that controls the movements of each read/write head in a hard drive. Heat from the drive can cause elements in the head stack to expand, putting the drive out of alignment.
"Freezing the drive could theoretically return the head stack to its original orientation," Reinert said. "But the modern servo [the motor that controls the head movements] is much better at making self-adjustments. Drives from 15 years ago used dedicated servos, where only one of the heads on a head stack was used to align the entire head stack, so the stack was very vulnerable to even microscopic changes in orientation. Older drives typically used 12-20 data surfaces and their physical dimensions were much bigger, so even slight alignment differences were exaggerated, compared to today's models that rarely use more than six surfaces, and are much, much smaller."
3. Stiction. This term means the drive's heads adhere to the surface of the disk platters. Normally, the heads fly over the platters on a cushion of air generated by the movement of the disk platters, but when the drive is shut down, the heads come to rest in an area of the platters designated for parking the heads, known as the "landing zone."
"Occasionally when the platters stop spinning, the heads may stick to the platter surface," Reinert said. "Freezing a drive may get the heads unstuck from the platters, although it's more likely to make the situation worse by damaging the delicate surfaces of the heads."
In summary, freezing a drive is like trying to kill a fly with a sledgehammer. It might work, but who knows what damage you'll do in the process? Plus, it's usually easier to just use a flyswatter.
If the data on the drive is worth getting back, no price will be too high to pay for a professionally-engineered recovery job. That being said, if you have a dead drive on your hands and want to try this technique (no guarantees!), here's a step-by-step guide.
Step-by-step guide to freezing a hard drive
- Seal the drive in a Ziploc bag. This prevents condensation from accumulating inside the drive and turning into ice. Contrary to conventional wisdom, hard drives are not vacuum-sealed. They have breathing holes with filters to keep dust out, so condensation can accumulate inside a drive pretty easily. Some people wrap the drive in a layer of paper towels as a further guard against condensation before sealing the bag up. Be sure to squeeze as much air out of the bag as possible.
- Leave the sealed drive in a conventional household freezer for about 24 hours. This allows the drive's temperature to be fully equalized.
- After a day or so, remove the drive, connect it to the PC and use some data recovery application to see if the drive is readable. Speed is essential -- copy everything out as quickly as you can.
- If the drive begins to fail again, make a note of where it began to fail, re-chill it and continue where you left off.
You could also place a set of freezer-chilled cold packs—the kind doctors give you for a sprained wrist -- in the freezer along with the drive when you first cool it down. Wrap the drive in the cold packs when you take it out. This allows the drive to remain cool longer than if you were simply running it out in the air.
Other ways to keep the hard drive cold
Other ways to keep the drive cool include a bed of dry ice in a styrofoam cooler. Apparently the cold-pack trick also works for notebooks where the hard drive can be exposed by opening a service door. Place the cold pack against the drive to keep it cooler that much longer. Don't attempt to flash-freeze the drive or to chill it far below the freezing point of water, which may only make things worse. You could crack the platters or cause other untold damage.
Some people have had success hooking up the drive to an external USB or Firewire cage and placing that in a bag, then in the freezer, and running the data and power wires out of the freezer through the door seal. This way the drive can theoretically be run indefinitely without having to re-freeze it.
This isn't a bad idea, but some drives with hard errors may behave strangely when mounted in an external (USB/Firewire) chassis. Some such drive cages don't correctly report hard errors across the USB bus, so the copy operating may simply "go off into space" or fail for reasons that can be ameliorated by mounting the drive directly in a PC.
To that end, for the best possible recovery operation you may need to mount the drive directly on the IDE or SATA bus. (One truly wild idea involves running SATA/IDE cables and Molex power connectors into the freezer as well, which is entirely feasible as long as you don't mind it being a bit awkward!)
Finally, let me emphasize: This freezing trick is not a repair. It's just a way to get the drive running long enough to remove the data and migrate it to something that works. If the data on that drive is irreplaceable, your best bet is always to call a professional data recovery service such as DriveSavers or OnTrack, instead of trying something that's, at best, a roll of the digital dice.
About the author: Serdar Yegulalp is editor of the Windows Insight, (formerly the Windows Power Users Newsletter), a blog site devoted to hints, tips, tricks and news for users and administrators of Windows NT, Windows 2000, Windows XP, Windows Server 2003 and Vista. He has more than 12 years of Windows experience under his belt, and contributes regularly to SearchWinComputing.com and SearchSQLServer.com.
