Hrrm. I've had a similar experience but with opposite outcome.
Edit: There's also the trick about reseating chips by warming the motherboard in the oven, or something along those lines. Never could get my parents to let me try that one out.
Never seen this done to a whole laptop, but it wasn't an uncommon last ditch method for pulling files off a dying HDD back in the days of spinning platters.
Preferably in a zip lock vacuum bag and left over night.
This is a 17 year old article, but I'm a little skeptical that 10 minutes in the fridge whilst still in the laptop would have made much difference. Probably the forced reboot did just as much.
Reminds me of a time that a friend of a friend had me attempt to reflow an old laptop with a $20 heat gun. HP DV6000 or a 9000, if anybody remembers those failure rates. He had two.
I'd worked in a warranty depot and had some experience with the tear down by then.
The reflow didn't work for us, but we did manage to make one of the two functional again.
I once had an issue with my LG5 phone who entered in a boot-loop (also a common issue with LG4). I don't have a head gun, so I put it in the mainboard of the cell phone in oven for maybe 10 min at 150°C to reflow the soldering. At least I could reboot the cell phone and copy the important files. But after about an hour, it enter again the in boot loop and was broken for good.
I normally solder at 300C-350C, I can't imagine under 200 reflowing the solder, especially without contact. I have a cheap hot air rework station that I run at 400C-450C (or so it claims) and it is still quite difficult to get anything molten with it (except nearby plastic that I forgot to cover in kapton).
Maybe the closed space has some advantages that make up for the lower temp.
As to the fridge point, fridge vs. room temp probably wouldn't make much difference in what I glean of the scenario from a quick glance. It'd help a bit - would cool things a bit faster,* but, eh ... I wouldn't expect that to necessarily be a KEY element.
That written, temperature may still have been quite important (i.e., reboot alone wouldn't help). Particularly with a device containing something like spinning platters (as well as potentially extra heat generation associated with certain failure modes), and the additional factor of laptop heat load trade-offs, ... simply having the laptop off for some minutes might have been essential. In fact, the writer notes that transfer stopped during the first attempt to retrieve the data (after 10 minutes in fridge), hence a second round of 20 minutes was carried out after which the remaining data was retrieved.
The fact that that second pause was 20 minutes AND was carried out in a fridge may have had little to no role in success the second time. To me, that's something like a 50-50 odds event ... particularly with this kind of infrequent (for an individual, hopefully) 'high-terror' n=1 event / 'data'.** It may have primarily been the difference between having transferred other files prior to transferring that (last? not checking that detail) file vs. not.
* Newton said proportional to temperature difference, but that's a rather rough guide in general for this sort of scenario - even with just basic details missing here
** I.e., where most people aren't too interested in carrying out any kind of more detailed post-mortem ... most are much more interested in leaving the terror part behind.
Edit: OTOH, one should not discount the power of belief out-of-hand - i.e., 'externum-placebo' or something to that effect, if you please (my verbum fabulists will know). As a certain purveyor of cheap movie-branded trinkets once said in response to the soft and breathless exclamation of "I don't believe it!": "That ... is why you fail."
(Though I may be mixing up my little Gs from movies ...)
In any case, the fridge may have had more of an effect than my crudely rational analysis above suggests. {In any case}^2, this entirely unnecessary but necessary addendum will conclude with - ha ha, but, also, serious.
I can say from experiencing, 20 mins in the freezer for a bare 3.5 or 2.5 HDD would give about 5 mins of "quick, find important stuff" time...
>_last ditch method for pulling files off a dying HDD_
Before pulling it apart and and putting the platters in an identical donor drive. (Dam IBM DEATHSTAR drives back in the day!)
>I can say from experiencing, 20 mins in the freezer for a bare 3.5 or 2.5 HDD would give about 5 mins of "quick, find important stuff" time...
Well that's just pulled up a few traumas I'd forgot I'd buried. Gawd, that repeated "race against time" of that 5 minutes when it's working. Knowing that every time you do the freezer/recover cycle, it's closer to straight-up never working again.
I have also tried the Laptop baking many years ago with an ancient Dell Laptop which was known for its buggy internal Nvidia GPU.
However, I have learned in the meantime that lead-free solder requires temperatures much higher than achieved by baking PCBs in a literal kitchen oven.
Is there an explanation why this method is nevertheless so popular (maybe even successfull)?
The PS3 issue described in the video was more specific to it, but similar in it's nature.
TLDR from my very imperfect memory:
The whole board-solder-chip sandwich expands at different rates due to heat, so internal stress and cracks develop in solder joints, which makes them fail. Baking PCB in an oven does something like making the crack close a bit, so that the solder joint works again.
Problem is, neither the crack, nor the internal stress that has been built up is gone, so the solution actually doesn't last.
I remember getting my 9800 GT GPU to work a bit longer this way.
I was playing Mount & Blade back then, but noticed that the polygons would glitch out (e.g. vertices in random locations instead of where they should have been, like stretched out 3D models), shortly after which there would be screen artifacts (random colored pixels) and it would all freeze. After that, the PC would fail to boot.
I took the GPU out, propped it up on some tin foil balls, put it into the oven on over 210 C for a while and after that it would start working again. Seems like something was wrong with the card though, because eventually the temperatures would get to 90-100 C and it would crash, but until that point I put it in the oven like 3-4 times and it seemed to work for a little bit every single time.
After that I replaced it with a GTX 650, which still runs to this day in a now friend's computer.
When I was ~11ish I found my desktop was overheating quite regularly. The solution I came up with on my $0 dollar budget was to pop the side of the case off and point a box fan into it. Worked like a charm…for a while.
I did this a lot as a kid, up to the point where I just didn’t care about having a case anymore. I completely removed the whole thing and put it on top of a rubber plate, and covered it after use with an old horizontal metal case from a 486 or something. I loved how confused my mother and then later my girlfriend would look when they saw me working/gaming in that total barebones setup. To be honest I have very fond memories of it and would definitely do it again if I didn’t have two small kids!
My current desktop setup uses a mining rig frame and tends to get the same confused looks from people.
I have it setup that way because there are no motherboard options for my CPU which allow for sufficient spacing between two triple slot GPUs to provide adequate airflow.
Yup. I tried this on a whim on a self assembled AMD K6 and it worked. Without that 'fix', compiling the Linux kernel would fail. Compiling the kernel and making GNU Chess play itself at highest levels were my standard stress tests.
This box wasnt completely self assembled in the sense the CPU was already mounted on the motherboard. Redoing the heat sink with thermal paste fixed the problem, mostly.
There was also the opposite direction: Putting a drive that's locked up in an oven on low heat to unstick the drive-head.... I knew someone who had to do that ~30 years ago after a long-running server had a power failure and the head had stuck to the platter as it cooled.
This was back when opening the drive itself up to figure out what was wrong was still reasonably unlikely to make things worse (I know that first hand from having "kickstarted" my own drive every morning for months while saving up for a new one by opening it up and nudging the platter)
Ive recovered a few drives through the freezer trick. It took a lot though. Freezer for an hour, use until lock, freezer for two hours, then rinse and repeat until data recovered completely. Doesnt have to be overnight, just long enough to get the metal to contract, and theres a point where that change no longer happens and keeping it in the freezer is only wasting time.
On the topic of reflowing in your home oven.. **DONT**!!!!! Thats how you get metal poisoning or other non fun diseases. If you plan on reflowing, use an oven that can go to the scrap yard after, or that will be dedicated solely to reflow. Reflowing can cause vaporization of metals in the solder and the chemicals from flux, etc, and thats just not fun to eat.
You're absolutely right about the health risks. Additionally, a small oven can be bought online for under $/£/€50 in most countries, and there are many guides online about modifying these to have precise temperature control for delicate soldering jobs. These modifications use inexpensive (usually <€1 each) thermocouples, along with digitally controlled relays or MOSFETs to do pulse-width modulation.
Reallyt didn't think it would work. And when it acctually did, I was sure it would only be a temporary solution. Much to my surprise, it's still going strong today (mostly used as an Octoprint server for my 3d printer.)
Back in the day, our IT guy in his 70’s pulled this trick on someone’s hard drive to recover a critical last minute presentation. The story became a staple of office parties. See you on the other side John.
I tried resoldering a laptop in an oven once in 2007ish. Sadly it only worked a very short time, like, maybe got the boot screen (I don't quite remember) and it uh was not in good shape. Dumb things you do in college
I had an iPhone 4S that would often lose wifi connectivity. But if I cooled it down in the fridge for a while it would stay connected for longer.
(Would want it off when in the fridge otherwise the battery life would get obliterated. Presumably from looking for signal but it did also sometimes die abruptly on very cold winter days.)
I had an old Galaxy smartphone that had some kind of thermal issue. It would constantly run hot, and, god forbid it should run out of batteries, would only start after it had been lightly chilled in the freezer. When it was running it performed okay but I ended up replacing it because it wasn't exactly reliable in an emergency situation.
Smaller world, I also used to work with him. And just recently used the fridge trick on an old dying hard drive (it worked fwiw), but don't remember if he was the source of why I thought that might work.
I've done this once. I got a laptop from a new company that was really shitty, but I had hopes that running Gentoo with some nasty CFLAGS would make it usable. However, the thing would heat a lot when compiling the base system and would crash in the middle of the installation.
Fair enough! I put it in the fridge and ran the installation from there. It concluded without issues and the laptop was MUCH faster than with a "normal" Linux distro. (I don't think this kind of big performance difference holds up on modern machines though).
I guess that is meaningful in a location with higher ambient temperatures. Here we just open up the computer a bit and aim a big table-fan at it for better cooling :-)
Well you know the old adage : Gentoo is compiled for your computer in its condition so that’s why it’s fast. Here the Gentoo was baked in excellent atmospheric condition so that’s why it was an awesome install, configured for coolness :)
I think it was Debian with i386 packages. Compiling with Gentoo targeting my CPU and not a generic i386 machine made things much faster for such a crappy device.
It's just that for recent distributions, if you have a reasonably common and modern CPU, you already get most of the optimizations you need in the pre-compiled packages.
When my laptop was getting extremely overheated during certain intensive computation (not necessary having drive issues) I used to grab a bag of frozen peas from the freezer, put an extra ziplock around it to prevent moisture, and pop it under my computer. Worked like a charm to bring the temp down and get through the crunch.
- HP laserprinter motherboards, 2 of them, one just a few days out of warranty. After a few minutes in the oven they went on to work for years and might still do so for all I know.
- HP JetDirect cards, 3 of them. All worked fine after baking, one of them needed a second bake after about a year. One of those is still in use, the other 2 got fried by lightning strikes - that was before I installed a surge protector in front of the fuse box.
- HP DV6000 motherboard, it worked for a number of years after that and probably still does, the LCD display eventually died which was the end for that piece of plasticky garbage
Notice a trend? HP seems to have had problems during the introduction of the RoHS directive [1].
- Asus something-or-other laptop motherboard, still works, one of the SoDIMMs died which left it with only 2 GB and it got retired
- PSU board for a 24" Hyundai monitor, still works and is hooked up as second monitor to the...
- Apple "Late 2009 27" iMac" graphics card, still works after the first bake, I'm using that machine (running Debian) as my main workstation.
The oven and the BGA reworkstation have saved a lot of equipment from early retirement, both at places I worked as well as here. I just repaired the PSU board for a HP (again HP...) 2910ag switch even though I could have swapped it on warranty because replacing a transistor and resistor is much quicker than going through the motions of a warranty replacement and to be honest also just because I can and to be even more honest because I only found out about the 100 year warranty on these switches after I had already fixed it...
>Notice a trend? HP seems to have had problems during the introduction of the RoHS directive [1].
It wasn't just them. I got an old Sony PVM-8044Q from an auction a year or two ago that was stuck displaying black and white on one of its inputs. Apparently a known[0] problem. I took off the case and located the general area of the PCB said to have issues and reflowed the solder of a few points with my iron. Color worked after that!
I have done this. Many years ago I was forced to work with a terrible laptop. It was one of those Dell 5cm thick things that were basically cramming a desktop cpu and gpu into someting that looked like someone had bolted a second box undeneeth a standard laptop. Fans ofc screeming loudly all day. So I had the brilliant idea to just run everyting of an external drive so that when I could work from home, I could seamlessly switch to a better computer.
Now the powerbrick for the laptop from hell was also huge and heavy (I remember the total package was over 4.5 kg, and the battery lasted 80 minutes when new, so you always needed wallpower), but it had the exact with and lenght of the exteral drive cage. So fancy me had the great idea that when I tidied up my desk putting the drive on top of the powerbrick, ofc neatly allingned with the desk edges, looked perfect.
Perfect until the drive crashed hard due to being right on top of the very hot powerbrick. All code was checked into subversion, but the 127 pages of a product manual I had been writing the last week were gone without backup.
Having come across the freezer method, I decided to give it a try. Nicely sealed in a ziplock freezer bag, I left the drive in overnight. Tried booting it up the next morning without success. Put it back in the freezer on a whim, but had to start writing the manual anew and forgot a out the drive.
I came across the drive again while fetching some peas from the freezer a few weeks later. Gave it another try, and lo and behold, it worked!
Learned my lesson about excess heat on drives and no nightly backups though, so never needed this procedure again.
Once I worked in a small office and we used ancient but silent servers and we needed new ones. We had a rack in one of the rooms, so without thinking I bought two new rack servers. Worst decision ever.
They were screaming loud. You could hear them in all the offices. Temperature in that room went up to 45°C. In the end we could store them somewhere in a closet and ignore the noise.
The massive noise difference between brands, models, and generations of rack servers is astounding to me. I have five rack servers in my house, not in a closet. Three are Dell R620s, one is a Supermicro X9, and one is a Supermicro X11. The Dells are 1Us and by all rights should be screamers, but they aren't. The Supermicros are 2Us, and while they weren't screamers, they were loud enough that on the one that's on 24/7 (the X11), I swapped the PSUs for quieter models, the CPU fan for a Noctua, and throttled the case fans. When the X9 boots up to ingest ZFS snapshots, you can definitely tell it's there.
I used to think the R620s being bizarrely quiet was because homelab loads are understandably small, but then one day I got well over half the cores to 100% sustained doing something, and it didn't get any louder. I know they _can_ get loud, because at boot the fans go to 100%.
Pro tip: Install the cables on the drive before putting it in the freezer inside the ziplock bag. Close the bag with duct tape or something to stop air getting in. Then when you take it out to use it, keep the drive in the bag and just plug in the cables. Keeping the frozen drive in the bag will prevent the instant condensation the moment you remove it from the freezer.
Double pro tip (I've done this): Forget the freezer. Put the drive a bag, with cables hanging out. Then immerse the drive/bag in a bucket of ice water. That keeps it cool even while being used. Just be careful to keep the cables away from the water. Last time I did this I used a garbage bag large enough that I didn't need to seal anything.
I think this is genius tbh. If you're faced with the prospect of losing important data and feeling silly for trying something so unorthodox I know which one I'd prefer. I've also seen videos on YouTube where repair technicians are able to temporarily fix broken (clicking) hard drives by giving them a good flick with their finger. It seems like magic to me. I've had a hard drive fail recently and I wish I knew about these methods. I wonder if there's any other hacks that might work for saving failing drives?
176 comments
[ 3.7 ms ] story [ 102 ms ] threadI will be trying this if I ever experience it. I don't have computer hardware fail that often.
Try copying files, fails. Try copying files, fails. Try copying files, fails.
Put laptop / drive in fridge. Try copying files, fails. Works.
"Hey, putting it in the fridge worked!"
Nope..it worked because they kept trying. People who comment on things like components shrinking don't understand how hard drives work.
Edit: There's also the trick about reseating chips by warming the motherboard in the oven, or something along those lines. Never could get my parents to let me try that one out.
in order to get everything off the phone i froze it in freezer. it gave me enough time to boot it and copy everything via usb
Preferably in a zip lock vacuum bag and left over night.
This is a 17 year old article, but I'm a little skeptical that 10 minutes in the fridge whilst still in the laptop would have made much difference. Probably the forced reboot did just as much.
On the other end of the heat spectrum, there's baking a motherboard to reflow the solder: https://www.reddit.com/r/talesfromtechsupport/comments/6dygu...
I'd worked in a warranty depot and had some experience with the tear down by then.
The reflow didn't work for us, but we did manage to make one of the two functional again.
That written, temperature may still have been quite important (i.e., reboot alone wouldn't help). Particularly with a device containing something like spinning platters (as well as potentially extra heat generation associated with certain failure modes), and the additional factor of laptop heat load trade-offs, ... simply having the laptop off for some minutes might have been essential. In fact, the writer notes that transfer stopped during the first attempt to retrieve the data (after 10 minutes in fridge), hence a second round of 20 minutes was carried out after which the remaining data was retrieved.
The fact that that second pause was 20 minutes AND was carried out in a fridge may have had little to no role in success the second time. To me, that's something like a 50-50 odds event ... particularly with this kind of infrequent (for an individual, hopefully) 'high-terror' n=1 event / 'data'.** It may have primarily been the difference between having transferred other files prior to transferring that (last? not checking that detail) file vs. not.
* Newton said proportional to temperature difference, but that's a rather rough guide in general for this sort of scenario - even with just basic details missing here
** I.e., where most people aren't too interested in carrying out any kind of more detailed post-mortem ... most are much more interested in leaving the terror part behind.
Edit: OTOH, one should not discount the power of belief out-of-hand - i.e., 'externum-placebo' or something to that effect, if you please (my verbum fabulists will know). As a certain purveyor of cheap movie-branded trinkets once said in response to the soft and breathless exclamation of "I don't believe it!": "That ... is why you fail."
(Though I may be mixing up my little Gs from movies ...)
In any case, the fridge may have had more of an effect than my crudely rational analysis above suggests. {In any case}^2, this entirely unnecessary but necessary addendum will conclude with - ha ha, but, also, serious.
I can say from experiencing, 20 mins in the freezer for a bare 3.5 or 2.5 HDD would give about 5 mins of "quick, find important stuff" time...
>_last ditch method for pulling files off a dying HDD_ Before pulling it apart and and putting the platters in an identical donor drive. (Dam IBM DEATHSTAR drives back in the day!)
Well that's just pulled up a few traumas I'd forgot I'd buried. Gawd, that repeated "race against time" of that 5 minutes when it's working. Knowing that every time you do the freezer/recover cycle, it's closer to straight-up never working again.
However, I have learned in the meantime that lead-free solder requires temperatures much higher than achieved by baking PCBs in a literal kitchen oven.
Is there an explanation why this method is nevertheless so popular (maybe even successfull)?
The PS3 issue described in the video was more specific to it, but similar in it's nature.
TLDR from my very imperfect memory: The whole board-solder-chip sandwich expands at different rates due to heat, so internal stress and cracks develop in solder joints, which makes them fail. Baking PCB in an oven does something like making the crack close a bit, so that the solder joint works again.
Problem is, neither the crack, nor the internal stress that has been built up is gone, so the solution actually doesn't last.
I was playing Mount & Blade back then, but noticed that the polygons would glitch out (e.g. vertices in random locations instead of where they should have been, like stretched out 3D models), shortly after which there would be screen artifacts (random colored pixels) and it would all freeze. After that, the PC would fail to boot.
I took the GPU out, propped it up on some tin foil balls, put it into the oven on over 210 C for a while and after that it would start working again. Seems like something was wrong with the card though, because eventually the temperatures would get to 90-100 C and it would crash, but until that point I put it in the oven like 3-4 times and it seemed to work for a little bit every single time.
After that I replaced it with a GTX 650, which still runs to this day in a now friend's computer.
I have it setup that way because there are no motherboard options for my CPU which allow for sufficient spacing between two triple slot GPUs to provide adequate airflow.
This box wasnt completely self assembled in the sense the CPU was already mounted on the motherboard. Redoing the heat sink with thermal paste fixed the problem, mostly.
This was back when opening the drive itself up to figure out what was wrong was still reasonably unlikely to make things worse (I know that first hand from having "kickstarted" my own drive every morning for months while saving up for a new one by opening it up and nudging the platter)
On the topic of reflowing in your home oven.. **DONT**!!!!! Thats how you get metal poisoning or other non fun diseases. If you plan on reflowing, use an oven that can go to the scrap yard after, or that will be dedicated solely to reflow. Reflowing can cause vaporization of metals in the solder and the chemicals from flux, etc, and thats just not fun to eat.
Reallyt didn't think it would work. And when it acctually did, I was sure it would only be a temporary solution. Much to my surprise, it's still going strong today (mostly used as an Octoprint server for my 3d printer.)
Otherwise you would have really had to get a dedicated oven to do that reflowing. Or lead all over your oven.
I reckon across 3 boxes I did this to, I eked out a couple hundred more hours of use.
Intel MacBook Pro combined with Microsoft teams meetings does this all the time for me.
I have a gel ice pack in the frezer.
When I put it under the laptop it cools down very quickly and I can work again.
(Would want it off when in the fridge otherwise the battery life would get obliterated. Presumably from looking for signal but it did also sometimes die abruptly on very cold winter days.)
If you're reading this Adam, they never got back to me on the W2. I ended up having to call the corporate headquarters.
I'd expect this to be a less useful trick these days. Not sure.
Fair enough! I put it in the fridge and ran the installation from there. It concluded without issues and the laptop was MUCH faster than with a "normal" Linux distro. (I don't think this kind of big performance difference holds up on modern machines though).
So you need to worry about condensation, when you take the equipment out of the fridge at the end.
(Where I live now, our dew point is typically at 24C throughout the day. So anything at or below that temperature collects water. That's about 75F.)
Do you mean one that ships binaries, or perhaps one that also enables a GUI by default?
It's just that for recent distributions, if you have a reasonably common and modern CPU, you already get most of the optimizations you need in the pre-compiled packages.
Ice packs on top of the routers.
I thought these were common practices all along. :-)
- HP laserprinter motherboards, 2 of them, one just a few days out of warranty. After a few minutes in the oven they went on to work for years and might still do so for all I know.
- HP JetDirect cards, 3 of them. All worked fine after baking, one of them needed a second bake after about a year. One of those is still in use, the other 2 got fried by lightning strikes - that was before I installed a surge protector in front of the fuse box.
- HP DV6000 motherboard, it worked for a number of years after that and probably still does, the LCD display eventually died which was the end for that piece of plasticky garbage
Notice a trend? HP seems to have had problems during the introduction of the RoHS directive [1].
- Asus something-or-other laptop motherboard, still works, one of the SoDIMMs died which left it with only 2 GB and it got retired
- PSU board for a 24" Hyundai monitor, still works and is hooked up as second monitor to the...
- Apple "Late 2009 27" iMac" graphics card, still works after the first bake, I'm using that machine (running Debian) as my main workstation.
The oven and the BGA reworkstation have saved a lot of equipment from early retirement, both at places I worked as well as here. I just repaired the PSU board for a HP (again HP...) 2910ag switch even though I could have swapped it on warranty because replacing a transistor and resistor is much quicker than going through the motions of a warranty replacement and to be honest also just because I can and to be even more honest because I only found out about the 100 year warranty on these switches after I had already fixed it...
[1] https://en.wikipedia.org/wiki/Restriction_of_Hazardous_Subst...
It wasn't just them. I got an old Sony PVM-8044Q from an auction a year or two ago that was stuck displaying black and white on one of its inputs. Apparently a known[0] problem. I took off the case and located the general area of the PCB said to have issues and reflowed the solder of a few points with my iron. Color worked after that!
[0] https://crtdatabase.com/crts/sony/sony-pvm-8044q
Ironically it was from a company called "Lava" so ...
Now the powerbrick for the laptop from hell was also huge and heavy (I remember the total package was over 4.5 kg, and the battery lasted 80 minutes when new, so you always needed wallpower), but it had the exact with and lenght of the exteral drive cage. So fancy me had the great idea that when I tidied up my desk putting the drive on top of the powerbrick, ofc neatly allingned with the desk edges, looked perfect.
Perfect until the drive crashed hard due to being right on top of the very hot powerbrick. All code was checked into subversion, but the 127 pages of a product manual I had been writing the last week were gone without backup.
Having come across the freezer method, I decided to give it a try. Nicely sealed in a ziplock freezer bag, I left the drive in overnight. Tried booting it up the next morning without success. Put it back in the freezer on a whim, but had to start writing the manual anew and forgot a out the drive.
I came across the drive again while fetching some peas from the freezer a few weeks later. Gave it another try, and lo and behold, it worked!
Learned my lesson about excess heat on drives and no nightly backups though, so never needed this procedure again.
This hurts just reading it. How'd it go in the end?
Once I worked in a small office and we used ancient but silent servers and we needed new ones. We had a rack in one of the rooms, so without thinking I bought two new rack servers. Worst decision ever.
They were screaming loud. You could hear them in all the offices. Temperature in that room went up to 45°C. In the end we could store them somewhere in a closet and ignore the noise.
I used to think the R620s being bizarrely quiet was because homelab loads are understandably small, but then one day I got well over half the cores to 100% sustained doing something, and it didn't get any louder. I know they _can_ get loud, because at boot the fans go to 100%.
Double pro tip (I've done this): Forget the freezer. Put the drive a bag, with cables hanging out. Then immerse the drive/bag in a bucket of ice water. That keeps it cool even while being used. Just be careful to keep the cables away from the water. Last time I did this I used a garbage bag large enough that I didn't need to seal anything.