Interesting story. I bought a hard drive from a garage sale, as I was about to wipe it I found over 4000 photos of this guy in his college days, photos like college parties, holidays, friends etc all dating back to 2003, also a copy of an old Counter Strike game. Interesting enough, I was surprised all the files were there and not corrupted from being turned on for the first time since the last file was written in 2004. I tracked down the guy and found out he lives in another state, currently in the process of reuniting him with the hard drive he has lost for over 15 years!
@@khairulhafidz15 I was cleaning a connector in a 80 gb Samsung, and I got distracted by my daughter, me dropped it, not on the ground, but simply on the bench. But it was enough to damage it. Fortunately, it was nothing special, or had any important data.
The last minute of the video is actually the most important of this wonderful video. There are reasons for early HDD failure, but if you avoid harsh treatment, hard drives are spectacularly reliable and long lived. I fished a 100GB 2.5 inch IDE out of a dead laptop that had been resting untouched in my garage for ten years and recovered every byte on it. Probably 40 GB of actual data. I simply don't shake or bake my drives and they simply keep on humming along inaudibly. HDD's are also spectacularly inexpensive. Jewels for a pittance. Unbelievable what you can get for fifty bucks. You can keep your modern SSD , with built-in failure in a couple of years, and I'll keep enjoying my backward, care-free existence. I could be wrong, but I would hazard the guess that consumer computing is unprepared for the massive wave of data loss to come, as immature back-up practices collide with SSD end of life in a couple of years. Looking for someone out there to correct me.
@@deus_ex_machina_ Yeah, but full backup is a major pain in the ass if you have multiple TB's of data you wanna keep safe... at least if you do it manually, 'cause you don't trust automated solutions.
I built my own desktop - and yeah, HDD temperature is a bigger deal than most people think. Overheating drives does happen. If you build your own system, make sure the drives can be properly ventilated. Even if your system is pre-built, make sure to regularly clear out the dust from the drives as well as the other components.
Gotta love the way my laptop uses an SSD and hard drive. If the HDD isn't being used, it won't spin it up at all, thus greatly reducing the run-time of the drive in my computer. It's currently sitting at 29C.
Does it really matter, aside from battery life? Typically a HDD in a secondary role will outlast a laptop, and the low temperate isn't all that significant, meaning it would have to be tens of degrees higher before you could expect that to reduce lifespan below another failure modem but again, heat = power consumption and it's great to have storage that is only using power when needed, particularly in a mobile, battery powered application.
Makes no sense. Every spinup of the HDD is worse than it running consistently for a long time. My NAS HDDs have 25000 hrs of runtime, still doing fine.
The only apparent omission in this excellent video is to introduce and define MTBF- Mean Time Before Failure. I find your work, positively outstanding.
Hi Chris! My fave HDDs are WD and I use Data Lifeguard Diagnostics! I also like HGST. My fave drives were Samsung, before Seagate bought them out. I have two 500 GB Samsung HDDs with around 12,000 hours on them (around 10 years old). 100% perfect. I use Crystal Disk Info and display HDD temps on the taskbar. On data fade, I have Betamax home video movie tapes from 1984 still perfect! Great videi as always! 👍
The head is not moved by a servomotor. What moves the head is the Lorentz force. This works by applying a small current to a coil in a magnetic field thus moving the head left to right way quicker and more precise than a servomotor could.
Thanks for this, I stand corrected. Though I believe the setup is still called a "servo mechanism" (which is what I should have said) even when it is based on a (voice) coil and not a traditional servo or stepper motor.
servo = "short for servomechanism or servomotor" Mr. ExplainingComputers usage is normal in electronics. Why it is a servo is beyond the scope of the video, although it is super interesting. Hard drives of today could not possible work without it. sre
@@kennethflorek8532 yep some "vintage" computer RUclips channels had some videos showing really old HDD with real indexed stepping motors (servos), some of then had shown that by adding some oil into the bearing shaft came alive long enough for data recovery.
@@RetroTinkerer The head actuator of less expensive hard drives in the era of the original IBM XT/AT is exactly what is not a servo. They are stepper motors that jump to align to the next coil each time they are powered. It is just assumed the head will be aligned with the track meant to read. (There is always some amount of slop, which is a problem.) What makes a head actuator a servo is the feedback to the motor derived from the track meant to be read, as it is read. The feedback is what pushes the head to align with the track. Stepper motors use "dead reckoning." Servos home in to the right spot. Hard drives in expensive mini-computes already had hard drives with servos, and so did some improved floppy drives, but the price had to come down to be usable for people who could only afford an IBM AT. The spindle motors (which turn the platters) are not stepper motors or servos. One of the 8 inch floppy drives I bought (before the IBM PC era) cost about $400, which is something like $1500 in today's money. The price of the cheap HD's for an IBM XT were (at first) around $2000, or $8000 in today''s money. That would be a lot cheaper than what a HD for a mini-computer would have been.
Excellent vid. One thing I've personally noticed is how much longer HDDs can store data long term. I retrieved files off an old IDE-type HDD that were over 20 years old! In comparison, I've had flash devices have their data get corrupted after around 5 years.
Well done, Dr. Barnatt! This new video of yours is incredibly concise, accurate, clear, practical, and enormously useful and valuable to those of us who use computers. I suppose we number more than a few these days. It comports exactly with what I've learned and consider important about disk drives after more than 44 years of experience with using different varieties of them. When I started, they were the size of automatic laundry-washing machines, and stored a few megabytes for thousands of dollars each. I was critical of your content last week, so I owe you high praise for a job well-done this time! Thank you, sir, and thank you for the valuable links as well.
Thanks for the video. Dell had a 50% dead-on-arrival rate for computers shipped to their customers in their early years. (Note: I edited and reworded this post without extreme sleep deprivation choosing my words for me.) The problem was largely due to a single cause, hard drive failures. Something had to be done. Dell invented a new term, the "hard drive touch". What a hard drive touch referred to was any time a computer hard drive was picked up, moved, vibrated, dropped, tossed, or bounced. Some of these words sound... crazy. Tossed? Why would hard drives be "tossed"? Keep in mind how Dell started. Dell helped pioneer the idea that computers could be manufactured to order, per the customer's specification. Then, the computers were only available via delivery, through the mail. During shipping and handling, Yes; often computers were dropped, tossed, or bounced Dell started at the hard drive manufactures and followed the hard drives all the way to final delivery counting the number of "touches" along the way. The total number of "touches" had to be greatly reduced. Conveyor belts, factories, fork lifts, and just about everything else in the supply chain all had to be redesigned. This was very expensive, but it paid off. Dramatically improved, dead-on-arrival statistics proved it. There was quite some delay, but it was Dell's investigations that lead a consortium of hard drive manufactures to add a new feature to all disk operated hard drives, "parked heads". One of the worst things that can happen to a drive, is that the read/write heads can bang into the disk's surface, damaging the read/write heads, and the data they are attempting to read. "parked heads" was a concept that when the drive is not reading or writing, the read/write heads should be safely parked off to the side, away from the disk's surface. This allowed for drives to be much more damage resilient. Hope this helps. If you enjoyed this comment, please do me two favors. One: click the like button on Chris's video above, and Two: Tune into my video comments 30 years from now, when I share a wonderful tale of how a little RUclips channel called Explaining Computers helped change the world for the better. : ) Again, thanks for the video.
Maybe I should power up my old machine from 2005 to see how it's doing... :) Been using it for 10 years before replacing it... with the same XP installation they've put on it in the shop!
Another great video! Thanks. They can also fail if you move the drive when it's spinning. As with any spinning disk if you tilt it on an axis perpendicular to the spin it can distort the platter and cause a head crash. It also causes extreme stress on the bearings because, as you learned in school, anything in motion wants to keep going in that same direction which in this case is a straight line (around the disk). The faster the drive is spinning the more force it requires to tilt the drive. When you apply the force the platters resist that motion and place the load on the bearings and the platters themselves causing them to warp. (side note: server hard drives running at 10K or 15K RPM are two or three times as thick as consumer drive platters and are usually much smaller in diameter. This is why "large capacity" SCSI drives are usually of less capacity than "large capacity" consumer drives.) If you take a spinning hard drive (a bad one..) that is running at full speed try flipping it on its side quickly and you will understand. You will feel the resistance to that movement and you will actually hear the motor on the drive slow down due to the extreme load on the motor and bearings and may actually hear that terrible squeak of the head contacting the platter surface. This and dropping the drive while it's running are leading causes of mechanical failure due to head crashes. Many of the newer drives on laptops have a G-Force detector that can move the head out of the way (park) the instant it detects that the machine is in free-fall.You can test this also by turning on your laptop and tossing it (carefully) into the air and then catching it. If your drive has this feature you should hear the CLICK of the head parking itself and possibly the drive will shut down. Do at your own risk of course.
I consider myself reasonably computer literate but I pretty much always learn something from your videos (including this one). Thank you for your efforts.
#1 cause of PC failures, for me, was motherboards, followed by optical drives, memory and finally, graphics cards and hard drives. Which is how I wound up with a heap of GPUs, HDDs, ODDs, PSUs and RAM; 7 (working) HDD made 2006-2010 total 830Gb, so their longevity is thwarted by improvements in capacity, sadly.
It's a blessing when SMART tells you that your hdd will gonna fail soon because you can backup your important data. Many hard-drive sometime just "suddenly dead" even though SMART monitoring software like crystal disk info or HD Sentinel tell us that hard drive are fine or 100% healthy.
Should be common knowledge by now to not assume that anything labeled as "smart" is smart. I have had a HDD and an SSD fail recently for unexpected reasons. One was spun up only for differential backup every couple days but had severe data deterioration and positioning issues at 10 years of age (masked because new writes from the backups didn't show any issue), and I had a Samsung SSD experience bit rot that the firmware is said to specifically avoid by refreshing old stored data.
@@Alfamoto8 two of my Seagate hdd died without any reason within a year of buying but another one which is named Toshiba is still up and running till now after two years of use...that's my story
One thing that never seems to get mentioned is if your hard drive is starting up in a cold place. like if you have it on a concrete floor. it can seriously damage the hard drive if its starting up cold. that same goes for sudden temperature changes if you open a window in winter, that will kill your fans too.
It never gets mentioned because there are few to no applications where people are running HDDs in non-climate-controlled areas on a concrete floor, and even then, the floor has nothing to do with it. If you just mean a cool basement floor in winter, no, that is nowhere near cool enough to cause a problem if the room is otherwise a hospitable (to humans) temperature, and opening a window in winter does nothing to my fans, in fact I have ventilation fans directly exposed to outdoor freezing temperatures. They merely need a lubrication that stays effective in the temp range of startup as well as operational temp.
@@stinkycheese804 You are talking from your own experience it seems. My experience is that it does matter because i live in a place where temperature changes of 20c to 40c can happen if i want some fresh air by opening a window, clearly not something you've had to think about. Computers are used in all sorts of places, not just at home on a desk. In the facebook data center where i work in Sweden, built here for the cold winters, we have to be very careful about temperature regulation as you may guess. May i recommend to you, look at googles study on hard drives so you can understand the lubrication is designed and proved to run best at 30 to 40c. temps below 10c can cause greater damage on start up, like anything mechanical, it will have an effect on its life span. The floor always matters in a cold room because heat raises... so naturally the coldest point is the floor..
Very good basic introduction to hard drives. I would add that reliability greatly increased at least twenty years ago, though I've had a fair share of failures. I have had some drives running continuously for up to thirteen years. A 1GB system drive ran on a Linux server from 1995 to 2008 (with one CPU/MB upgrade during that time, and various Linux system upgrades). After sitting in storage it still ran again in 2014. Another server (a Sun 386i) ran from 1988 for about ten years. One of the characteristics of Unix-like systems is that they frequently rearrange data (a kind of defragmenting), so fade is not that great a problem. Different story, of course, for the external backup drives. I still don't have any SSD's!
I find it amazing that you haven't talked about "disk metal extension" over time. Indeed for HDs that have been spinning 5-8 years (yes it's common on servers), the disk has a tendency of expanding. Thus if you have to turn off you disk long enough for it to cool down from its "nice and warm working temperature", it might shrink enough so the head becomes unaligned from it's previous format. Thus rendering the disk unreadable, and causing it to be unable to boot anymore. Rare but I've seen it twice in my life.
I have a Samsung 750GB SATAIII 3.5" drive that is now just over 10 years old been in 24/7 use for half that time or more. I still use it today, it runs every day, I save videos to watch later on it and use it for a temp drive when editing videos. Samsung has always served me well with their drives, that's all I use and this is why. Though this is the longest I've trusted a drive, I intend to use it until the end of it's life, still no bad sectors or heads and runs quite. It does use those fluid bearings I bet otherwise I would have been done long ago. Good video thanks
I think it's safe to say that most people don't know about fragmentation. Also, I'm liking the pacing of your videos. Makes for easy watching or listening.
After working in IT between '89 and '07, HDDs were poor in the early years but after about 2000 get very reliable. The computers would cycle through the system in about 5 years so I didn't get to see the long term reliability.
That was a point I raised. The thing is, the HDDs will last long enough as you will more than likely replace it due to lack of space before you ever see one fail. In comparison, SSDs are virtually guaranteed to fail. Also, due to how SSDs operate, you cannot securely erase data from them you do not wish to be recovered like you can on a HDD.
@@NeilRoy They aren't, the TBWs are just "guidelines", the cells might often still writes but performance will takes a turn for the worst as the TBW leaves the "safe" zone. So as long as you avoid the cheapos (QLC and TLCs based drives) I think you can get good mileages if not better than HDDs mileages out of them.
And yet, replace the windows OS and the data on the 'rotten bits' of another drive reappears and is easily accessible, it's only Windows that rots; everything slowly stops working, yet windows reports that everything is fine. I have DOS and OS/2 computers from 1992 and the data never suffered 'bit rot'. My 10 y/o Linux system still works great. Only Windows seems to suffer from the OS slowly degrading over time.
These might be good for stationary uses, but anything that moves more than occasionally I'd rather go with an SSD. Sure they cost a lot, but not as much as data recovery!
One does not have to worry about hard drive failure too much when using external drive that most of time is not running or powered up. That said, because electronics age. A hard drive used to store important data should be replaced with a new one before it gets passed the 10 year mark. Also super important data should be backed up to multiple devices, optical media and so on. Stored in right conditions and in a safe place. I find it incredible how little people back up data and just stor things in the cloud or just do not backup at all. The thing most people miss out on is that a backup archive or important data as mentioned in the video needs to be refreshed or maintained to avoid problems. Great video :-D
Also, regarding heat sensors at about 11:00, Windows 10 software updates can and will sometimes remove temperature sensor based fan speed handling. Yes, you can find a machine automatically updated and stopped running fans based on temperature - and simply running fans at idle. So if you don't hear the fans they might actually not be running anymore. You can find a program to retake control of the fans/sensors and fix this - hopefully before anything's cooked.
Excellent video sir. I didn't realise that memory fade would still occur even if the hard drive was regularly used or at least powered up. Although I now use SSDs in my computers, my NAS backups use hard drives, so I will check out the programs you suggest to refresh the data. Thanks for the tips. :-)
I don't know of many laptops that have HDDs these days (some of the real cheap ones do I guess) but I almost always recommend people spend the money and upgrade to an SSD (either buy a kit or have someone do the upgrade--although it's not very hard). This gives you the peace of mind of an SSD and more durability in case you drop your computer while it's running. Keep the old HDD as a backup just in case but I would recommend people use SSDs instead. For cooling, the best option is to put a fan in front of the drive, that pulls cool air in the front and exhausts it out the back (towards the CPU, if your drives are located in front of the mainboard/CPU). This also helps with cooling of other components. If you can't do that, or there is no bracket or place to mount a fan, consider getting a 5.25" drive bay cooler (some also work as a 3.5" drive bay adapter) and use that. IF that doesn't work, then the drive rail with heatpipes could help. Most aftermarket cases should have a fan bracket in the front, but many OEM/store bought computers--particularly the cheap ones) don't usually have a fan located in front of the drives. If you can find them, you may still be able to find a cooler with fans on the bottom that will blow air onto the circuitboard on the bottom, but I haven't seen these lately in computer stores (since most people are moving to SSDs or doing custom-builds with aftermarket cases that support HDD cooling via mounted fans in the case).
my main 500gb hard disk failed after almost 4 years . i lost incredibly valuable data . the issue was with the read head . either the servo motor or the whole head assembly was broken. i learned my lesson the hard way . please guys always backup your data
To serve as a confirmation bias to what you would have said 15 years ago, the only HDD I had fail was the one in my roughly 2 year old iMac. So your statement from 15 years ago still isn't very far off! :)
Thank you for the information regarding "data fade". I had no idea about this. It would totally suck to lose my backup data. Perhaps DVDs would do better for long time backup storage
Thanks for another lovely clear fact-filled video! Perhaps it’s because I’m English but I find your explanations much clearer than many from over the pond.
Like he said, it happens for every type of magnetic storage. This obviously includes video tapes and K-7 audio tapes as well, that's also something to keep in mind.
I have drives that have been stored in my un insulated garage for 20 years these recently were checked if they still functioned. They still had data intact and no bad sectors reported after a scan and were able to boot linux and dos 6.22. Higher density drives may be more subject to this than older drives. Scarily enough I still have Commodore 64 floppy disks from the 80's that are still readable. Magnetic storage such as floppies and tape storage are more of a worry for data fade than hard drives as they are less shielded that hard drives. I am more worried about my data stored on cheap optical media ie CD's from 20 years ago are becoming more unreliable depending on the storage environment and can be subject to delamination.
Tech Truth it's crap information I have a hard drive that's 25 years Old when I got it it still had all the stuff and read with no problems But the hardrive was only 200mb
Additionally, they dont make them like they used too. (As they say). I work in a DC, and we swap out maybe 30 to 40 dead drives every single week. The new servers with 750GB 15k drives tend to fail after about 1.5yrs. However for specific reasons we have some really old compaq ml370 servers with 9GB 15k drives that are still working fine to this day. Some 17yrs later !.
I don't like Linus' "american reality showmanship". Everytime I watch a video from that guy I get upset and stressed. Chris' videos causes me the opposite effect though. I feel relaxed and ending up with more knowledge that I begin with.
For me, this was one of your most informative videos yet. As a bench tech I'm always challenged when explaining to a casual user why a seemingly good drive should be replaced based on SMART data, you've given me some great information to pass on in simple terms. I'd be curious as to your opinion on the accuracy of some of these utilities that go as far as predicting the approximate number of days until failure. I use hard drive sentinel and on several occasions I have encountered drives that allegedly have days left. For fun I've run these drives in test machines and of course they don't fail on cue even months after they've been reported to expire.
Thanks for this kind feedback. I don't think any utility can predict the number of days to failure! Rather, the best that can be reported (or taken from such days-to-go predictions) is that a drive has a reached the point where it has a high probability of failure, and should hence be retired from critical use.
Thank you! Data Fade prevention is something I had not thought about! Just downloading DriveDX Trial version as a Mac alternative and will spend the day going through my drives and back-ups - thanks!
SMART was originally intended to catch predictable failures but in actual implementation it rarely catches anything until it's too late. This was apparently due to concern that such predictive information would likely increase in-warranty repair attempts and could provide necessary legal evidence for class action lawsuits. Instead of reporting failures before your data is lost the drive plays musical chairs with failed sectors until the unit has passed well beyond the warranty period and the defects begin to exceed the drive's ability to hide them. At which point it's time to buy yet another drive from Western Digital (which now includes Hitachi and Seagate).
When noodling around in the linux man pages, I saw some peculiar, forgotten, or ignored stuff. At one time, in the old days, the way to take care of bad sectors, which normally increased, was in the file system. It wasn't internal to the HD. I believe you can still have failing sectors marked for non-use, if you care to. Another thing: At the factory, before you get the HD, they map out bad spots, because all platters have them.
I bought a 1Tb Samsung HDD in 2011, used it for 4200 hours on my desktop, in 2013 I put it on a dvr, and in 2015 according to HD sentinel the performance was at 100% and the health at 0%, I didn't replaced it because I want to see when it will finally die, yes It is still alive and spinning 24/7, with more than 40000 hours, the model is HD103SJ, manufactured in 07/2010. By the way, the temperature inside that dvr is above 50°c, hard drives are great, I had an SSD and went back to an HDD. I like the channel.
Thorough and useful vid thank you. Wish I had this info a decade ago; could have saved a few HDDs. Iv ready somewhere that overdoing HDD cooling (below ambient) also reduces it life expectancy. (maybe due to condensation inside?) Iv been using HDsentinel for many years now. Iv also used reports from it to get a few disks under warranty replaced. Still i have lost 2-3 HDDs due to failures so far . Another very important cause of failure i have noticed is abrupt switch off of HDD due to any reason like system crash/loose contacts (ext. HDD). When this happens too often bad sectors accumulate over the disk and that very much reduces disk life. Apart from this, iv noticed, when the HDD hardware is apparently defect free the main culprit is temperature (esp. in laptops).
My computer is now over 10 years old (bought in August of 2010), but initially I only had a prebuild that would fail about 4.4 years later. Then I used a laptop for ~2.8 years, and after that period I finally built my own PC using working parts of my prebuild (CPU, RAM, mainboard, hard drive, DVD/CD drive) and bought the other parts online (case, power supply) since the ones of my prebuild were too small / weak for what I had planned. As for the graphics card, I was lucky and got a Zotac GTX 470 from a friend of my sister which I used for ~1 year although for some reason after a summer vacation (I had it for ~8 months already) it suddenly "died". However, I could still use it without drivers but it stuttered _a lot_ and I think sometimes it would even have visual artifacts on the screen. Well, needless to say I could barely play any games or even just watch videos / browse the web smoothly with that "zombie" of a card... Fast forward a few months in the autumn of 2018, I got my new graphics card at last! It is an MSI (AMD) Radeon RX 570 8 GB Armor OC - which I still use today. :) So this is the current state of affairs. Sorry btw, this turned out to be an essay lol
Wow thanks so much for this, I came here out of curiosity (I did know that HDDs lasted a while if cared for right) but I DID NOT know of "data fade" at all. Now I do and will be refreshing my backup drives at external locations 'refreshed" every year or so. It's mostly photos, but honestly they mean a lot so I want to keep them safe and backed up.
Having trouble booting up , ran a test and it says everything is ok except for the fan . Anyway won't let me start , boot up log on ,,,nothing . How can i recover the files , pictures , docs from my old laptop to my new laptop ???
Still here with my 80GB hard drive on my ThinkPad X200S. Though my Grandad still has his hard drive from the early 80s going in his really old laptop. It's quite impressive.
I only ever lost data to ONE failure of a HDD...but that was me being stupid and over voltageing it :( I Will most def get the disk fresh prog though since I pretty much do just shove my HDD into a cupboard. Lol
I'd never heard of data fade on a hard disk. Thank you, sir. I have been using a 2TB hard disk to hold media files, as well as a semi-regular backup image of my OS drive for a year now and will get refreshing software just in case, since a lot of the media files are permanent and probably never move.
Now WD green parks the head for you, my had over 600000 last time I checked. But yes, I herd that power failure was a dead sentenced to a spinning drive, at that time.
I don't know how common they are, but there are drives that have an acceleration sensor in them. When a drive falls it will detect this as a few milliseconds weightlessness and park the heads before the drive hits the floor.
Of particular interest to me was the concept of data fade, and refreshing your data to prevent that. I had never heard of that. Thanks! I'm also about to run the Data Lifeguard Diagnostic (after I close this browser window) :-)
I tend to dismiss "data fade", at least in the case of much older HDDs. For example, there are many videos on RUclips of people recovering very old computers, from the early 1980's, with hard drives ranging from 5 to 80 megabytes. These drives usually have sat, untouched, for many years. In a fair number of cases, the data is still accessible. In some, stiction, motor failure, or even a combination of both has made the drive useless, but that is by no means a guarantee of failure. Having said that, I partially agree with your recommendation to periodically refresh the data, though I would just regularly move the data to more modern capacity drives or formats.
I almost have the same HDD that is shown in the video that's just sitting in my drawer. It's the WD Caviar WD200, 20GB. I don't know how old is it but I know I connected it to an older computer something like 4 years ago and it worked, booted up Windows XP. Now I'm thinking to get a SATA-IDE adapter and connecting it to my PC to see does it still work. Great video, explained right to the point. I should really have more attention to my current HDD that's about 7 years old and possibly get a new one to be on the safe side. :)
HDDs in my experience are tough resilient beasts. I have a bunch of really old drives from the 1990s ranging from 10 to 80 GB that I have submitted to constant full erase/ write cycles performing back ups on a close to daily basis for 9 years now. Intending to test their durability I even have had them operate with poor cooling becoming really hot, reaching almost 70c and operating like that for hours. half of those drives were reported as imminent failure by monitoring software with less than 12 working hours left on them when I first got them. other than re magnetizing a handful of sectors via software on a couple drives None of them has given me any trouble and all have worked way longer than the predicted 12 remaining hours, even the "crappy" IBM travelstar drives.
Datafade is not a joke. I am mega paranoid about it, full backup refresh every 6 months. Just be sure to refresh backup disks one at time, use a surrogate disk as a temporary backup to your backup in case of failure during the refresh. If you really want to retain your data, have a distro stick to boot from during file copy operations with your backups. Dont use windows, Microsoft filed patents to compare your files to known hashes they are paid to look for and auto-delete them. Its still in the development stage, but it will eventually make it into the update pipeline, and I suspect they will not tell you. Plug your backup with all your movies and music into windows 10... and.. its gone! That's what they hope to roll out. Dont play games with your files, do data management and wrangling in linux.
it's only paranoia. Rewriting all the data every 6 months is straight up insane, even every 5 years makes no sense. In fact, it makes no sense to do that ever, since the drive can fail anyway. All you need to do is to check that the data is readable, nothing more. And thinking that Windows will read your every file, really? That would absolutely never happen. What's this patent you mention? I wasn't able to find anything related to hashing the contents of files.
If an HD starts ticking make immediately a backup, don't put your pc off!!! Or it's dead, In my 40 years been busy with computers it happened a lot. Although I had a few where I lost data, I always use enterprise HD's, but they also fail after an amount of time. Thanks for the clear tutorial.
Those rubber mounts for internal harddrives are to limit the noise produced by the drive vibrations, not to protect the drive itself. The absorption of these mounts is so limited that they won't help against falling or shocking.
If you fancy the DIY level of cooling you can use a aluminium/stainless steel kitchen pot scrubby pad and (if you use two drives on top of each other) piece of plastic to shield the drives PCB from shorting out. The scrubby will take away the heat from the top of the drive which then can be carried away by means of a fan or series of fans.
Years ago, back before many of you were born and before machines like the VIC-20, TRS-80 or the Commodore PET, the relatively non existent computer industry (there were big firms like IBM and Honeywell making huge mainframes but nothing for the home user) was engaged in an informal debate as to how to cheaply implement computer technology into homes, hospitals, elementary schools and small business. One of the initial topics was memory. Industry used large tape drives or punch cards to store programs and data. The fledgling home computer industry didn't want to go this route. Their idea was to store as much data as possible electronically without cards, tape or a very primitive optic drive. The problem was it couldn't be done. The cost was too prohibitive and the technology to cheaply manufacture high density solid state electronics was just not there. Technologies after the development of the transistor such the Op-AMP and the four NAND Gate chip were HUGE. So the industry had to turn to cassette tape drives, floppy disks and eventually CD/DVD burners and HDD hard drives (in early drives those disks were made from glass and weighed a "relative" ton causing many spinning motors to burn out after less than a year of use) . The SSD drive was where the consumer computer industry wanted to start and it took almost 50 years for it to get there. Today the HDD is on it's last legs. The detour is finally over. Already gone are punch cards, tape and floppy drives and almost gone is the DVD burner. The only reason the HDD is still readily sold is due to inventories and third world manufacturing costs. Once the inventories are gone so will be the HDD. For all my clients I am switching them over to SSD. Any HDD implementation is done externally. I am much older than the presenter of this video and I can say it has been a fascinating ride these past 63 years. When I started working in consumer home electronics a home computer was just an "Orwellian." concept. The Vacuum tube would still be king until 1970 and an ipod was a transistor radio with a crystal earphone. I will not live to see it but I believe some form of implantable computer technology (either biologically or electronically) will be dawning soon in the next 20 years
It's all about probability. Mine 2tb seagate with reputation of dying in months survived four years to first bad sectors and five years before it failed with ticking sound. You can't make any statement about reliability even if you have ten drives, cause you may be lucky or unlucky.
My Asus Republic of Gamers laptop came with a 2TB HGST drive as well as an m.2 SSD, but I kept the drive because it's super fast and very reliable. I added a 2nd 2TB Samsung drive as well, but it just hasn't got the same performance.
My WD drives usually go about 40,000 hours on them before bad clusters turn up or they start to get noisier or sound unusual. They usually still work fine at that point, but that's when I know it's time to replace them. At that many hours, the slightest sign of failure means there's a good chance they may fail soon & possibly dramatically. Loud & excessive clicking means failure is imminent.
Hard drive failure isn't planned obsolescence unless the manufacturer designs the drive to fail. In 95% of cases it's because of wear over time; the laws of physics and thermodynamics still apply, and that's something that can't be helped.
I always do a surface test (any software will do - I use Victoria, for example) of an HDD when I buy one - even if it's brand new. Because better safe than sorry. And one more interesting fact. The less you turn your HDD off/on - the healthier it will be (if it doesn't doesn't overheat from running 24/7, ofc - so use proper airflow). It's because of the expanding and shrinking due to it getting hotter then colder. Those micro-expansions will damage the actual hard disc inside the drive. And the motor (as any other motor out there) is most stressed when it spins up (starts up), so it also help to preserve it. That is an another reason I set the HDD spin-down timeout setting to 0 (disabling it) after Windows installation.
Bit of a sigh to open by saying HDDs fail for many reasons but SSDs only when reaching max specced writes. You even include manufacturing defects for the HDDs but don't consider them for SSDs. Also, rubber mounts make very little difference. They are not shock absorbers, so they can only absorb high frequency vibration, and even there you'd need more elaborate constructions to isolate the drive from the case, and usually you have to do that to shield the case from the drive's vibrations more than the other way round, because most harddrives vibrate noticeably. As for harddisk coolers, it is much better to pass on that (especially one with little surface area as the one shown) and instead put the drive in a fan air flow path. That alone should suffice for any situation, even just a silent fan. Very mild direct air flow is multiples better than none at all and still much better than passive air flow from case underpressure currents or such.
But SSDs are fairly likely to fail simply from component failures, which happen unexpectedly and can cause data corruption from mild to severe, or render the drive completely unrecoverable. While exceeding the write capacity is something you should be able anticipate, since you can easily check how many writes you've officially got left. And having an SSD stop working due to exceeding write capacity is not scary at all. All the data can easily be retrieved without problems. So one should only care about component failures, not at all about exceeding the write capacity.
anything that stores data should be designed to last as long as possible (what if you are working on some groundbreaking research and you lose EVERYTHING because your drive was designed to fail)
it should be designed to last as long as possible but that is not the mind set, its to have more of an guarantee of an "upgrade" cycle so you buy more drives more regularly, and freak people out and make them use raid 1 instead of 0 for their personal desktop that does have a backup
Andrew - possibly but not universally true and everything is subject to inevitable 'wear & tear'; but a wasted comment I'm afraid as the critical aspects are WHEN and UNDER WHAT CONDITIONS, (also, I guess, what can be done to delay the inevitable?)
Damn i wish i knew that hdds do eventually die. I had a pc that fried its motherboard so i stored the hdd. After a few years i got a new pc but my old hdd is not working :( I also had an external hdd with thousands of pictures and it failed too very recently. So if you have an hdd with your valuable data in it, do yourself a favor and make a backup.
For testing a drive, I use the “badblocks” utility that comes standard with Linux. It can do read tests and write tests, and also keep repeating the tests as long as new errors keep showing up. And it’s not vendor-specific--it works with *all* drives. SMART is a waste of time.
![VICTOR Denmark Open 2024 | An Se Young (KOR) [1] vs. Gregoria Mariska Tunjung (INA) [5] | SF](http://i.ytimg.com/vi/Y3QEyCih85s/mqdefault.jpg)
Interesting story. I bought a hard drive from a garage sale, as I was about to wipe it I found over 4000 photos of this guy in his college days, photos like college parties, holidays, friends etc all dating back to 2003, also a copy of an old Counter Strike game. Interesting enough, I was surprised all the files were there and not corrupted from being turned on for the first time since the last file was written in 2004.
I tracked down the guy and found out he lives in another state, currently in the process of reuniting him with the hard drive he has lost for over 15 years!
Great story indeed!
@@tamie341 he was virgin
any update on the story?
@@muhammadagungwicaksono4472 wkwkwkwkwk
Wow
It always amazes me how complex and sensitive hard drives are whilst not failing all the time.
I totally agree. It is amazing the accuracy they achieve.
Unless you accidentally drop one from a 30 cm height and it stops working
@@joseislanio8910 I once did that without knowing its fragility and accidentally broke them..... after that happened I'm super careful with hdds.
@@khairulhafidz15 I was cleaning a connector in a 80 gb Samsung, and I got distracted by my daughter, me dropped it, not on the ground, but simply on the bench. But it was enough to damage it. Fortunately, it was nothing special, or had any important data.
wow
The last minute of the video is actually the most important of this wonderful video. There are reasons for early HDD failure, but if you avoid harsh treatment, hard drives are spectacularly reliable and long lived. I fished a 100GB 2.5 inch IDE out of a dead laptop that had been resting untouched in my garage for ten years and recovered every byte on it. Probably 40 GB of actual data. I simply don't shake or bake my drives and they simply keep on humming along inaudibly.
HDD's are also spectacularly inexpensive. Jewels for a pittance. Unbelievable what you can get for fifty bucks.
You can keep your modern SSD , with built-in failure in a couple of years, and I'll keep enjoying my backward, care-free existence.
I could be wrong, but I would hazard the guess that consumer computing is unprepared for the massive wave of data loss to come, as immature back-up practices collide with SSD end of life in a couple of years. Looking for someone out there to correct me.
A very interesting point in your final paragraph. I agree that more and more people seem to be relying on single copies of files kept on one SSD . . .
exactly and all for that few tenth of seconds or minutes faster wait time.
HDD failure is one of my greatest fears.
same being a kpop fanboy i got so many fancams
For me its ssd
Backup your important data on multiple places (including at least one off-site) and ease your tensions.
@@deus_ex_machina_ Yeah, but full backup is a major pain in the ass if you have multiple TB's of data you wanna keep safe... at least if you do it manually, 'cause you don't trust automated solutions.
@@MetalTrabant robocopy makes it significantly easier
I built my own desktop - and yeah, HDD temperature is a bigger deal than most people think. Overheating drives does happen. If you build your own system, make sure the drives can be properly ventilated. Even if your system is pre-built, make sure to regularly clear out the dust from the drives as well as the other components.
Gotta love the way my laptop uses an SSD and hard drive. If the HDD isn't being used, it won't spin it up at all, thus greatly reducing the run-time of the drive in my computer. It's currently sitting at 29C.
Does it really matter, aside from battery life? Typically a HDD in a secondary role will outlast a laptop, and the low temperate isn't all that significant, meaning it would have to be tens of degrees higher before you could expect that to reduce lifespan below another failure modem but again, heat = power consumption and it's great to have storage that is only using power when needed, particularly in a mobile, battery powered application.
Makes no sense. Every spinup of the HDD is worse than it running consistently for a long time. My NAS HDDs have 25000 hrs of runtime, still doing fine.
The only apparent omission
in this excellent video is to introduce and define MTBF-
Mean Time Before Failure.
I find your work,
positively outstanding.
There are two kinds of Hard Drives in the world.
Drives that have failed and Drives that have not yet failed.
Totally true.
True but useless. The same is true for human and PC boards and your dog.
There are also 10 kinds of people, those who talk binary and those who don't.
01011001 01101111 01110101 00100000 01100001 01110010 01100101 00100000 01100011 01101111 01110010 01110010 01100101 01100011 01110100 00100000 01101101 01100001 01110100 01100101 00100000
+Free Saxon You are correct mate
Hi Chris! My fave HDDs are WD and I use Data Lifeguard Diagnostics! I also like HGST. My fave drives were Samsung, before Seagate bought them out. I have two 500 GB Samsung HDDs with around 12,000 hours on them (around 10 years old). 100% perfect. I use Crystal Disk Info and display HDD temps on the taskbar. On data fade, I have Betamax home video movie tapes from 1984 still perfect! Great videi as always! 👍
The head is not moved by a servomotor. What moves the head is the Lorentz force. This works by applying a small current to a coil in a magnetic field thus moving the head left to right way quicker and more precise than a servomotor could.
Thanks for this, I stand corrected. Though I believe the setup is still called a "servo mechanism" (which is what I should have said) even when it is based on a (voice) coil and not a traditional servo or stepper motor.
ExplainingComputers I really like your videos, quite informative and interesting to watch. Keep up the good work.
servo = "short for servomechanism or servomotor"
Mr. ExplainingComputers usage is normal in electronics. Why it is a servo is beyond the scope of the video, although it is super interesting. Hard drives of today could not possible work without it.
sre
@@kennethflorek8532 yep some "vintage" computer RUclips channels had some videos showing really old HDD with real indexed stepping motors (servos), some of then had shown that by adding some oil into the bearing shaft came alive long enough for data recovery.
@@RetroTinkerer The head actuator of less expensive hard drives in the era of the original IBM XT/AT is exactly what is not a servo. They are stepper motors that jump to align to the next coil each time they are powered. It is just assumed the head will be aligned with the track meant to read. (There is always some amount of slop, which is a problem.) What makes a head actuator a servo is the feedback to the motor derived from the track meant to be read, as it is read. The feedback is what pushes the head to align with the track. Stepper motors use "dead reckoning." Servos home in to the right spot. Hard drives in expensive mini-computes already had hard drives with servos, and so did some improved floppy drives, but the price had to come down to be usable for people who could only afford an IBM AT. The spindle motors (which turn the platters) are not stepper motors or servos. One of the 8 inch floppy drives I bought (before the IBM PC era) cost about $400, which is something like $1500 in today's money. The price of the cheap HD's for an IBM XT were (at first) around $2000, or $8000 in today''s money. That would be a lot cheaper than what a HD for a mini-computer would have been.
Excellent vid. One thing I've personally noticed is how much longer HDDs can store data long term. I retrieved files off an old IDE-type HDD that were over 20 years old! In comparison, I've had flash devices have their data get corrupted after around 5 years.
Can always count on learning something new when watching one of your videos, thanks!
My 20 meg hard drive that I bought in the 1990's finally failed last year, I recon thats pretty good going
Great going! :)
Well served! I guess it wasn't made in China back then...
that's*
Well done, Dr. Barnatt! This new video of yours is incredibly concise, accurate, clear, practical, and enormously useful and valuable to those of us who use computers. I suppose we number more than a few these days. It comports exactly with what I've learned and consider important about disk drives after more than 44 years of experience with using different varieties of them. When I started, they were the size of automatic laundry-washing machines, and stored a few megabytes for thousands of dollars each. I was critical of your content last week, so I owe you high praise for a job well-done this time! Thank you, sir, and thank you for the valuable links as well.
Thanks for the video. Dell had a 50% dead-on-arrival rate for computers shipped to their customers in their early years.
(Note: I edited and reworded this post without extreme sleep deprivation choosing my words for me.)
The problem was largely due to a single cause, hard drive failures. Something had to be done. Dell invented a new term, the "hard drive touch".
What a hard drive touch referred to was any time a computer hard drive was picked up, moved, vibrated, dropped, tossed, or bounced.
Some of these words sound... crazy. Tossed? Why would hard drives be "tossed"?
Keep in mind how Dell started. Dell helped pioneer the idea that computers could be manufactured to order, per the customer's specification. Then, the computers were only available via delivery, through the mail. During shipping and handling, Yes; often computers were dropped, tossed, or bounced
Dell started at the hard drive manufactures and followed the hard drives all the way to final delivery counting the number of "touches" along the way.
The total number of "touches" had to be greatly reduced. Conveyor belts, factories, fork lifts, and just about everything else in the supply chain all had to be redesigned.
This was very expensive, but it paid off. Dramatically improved, dead-on-arrival statistics proved it.
There was quite some delay, but it was Dell's investigations that lead a consortium of hard drive manufactures to add a new feature to all disk operated hard drives, "parked heads".
One of the worst things that can happen to a drive, is that the read/write heads can bang into the disk's surface, damaging the read/write heads, and the data they are attempting to read.
"parked heads" was a concept that when the drive is not reading or writing, the read/write heads should be safely parked off to the side, away from the disk's surface.
This allowed for drives to be much more damage resilient.
Hope this helps.
If you enjoyed this comment, please do me two favors. One: click the like button on Chris's video above, and Two: Tune into my video comments 30 years from now, when I share a wonderful tale of how a little RUclips channel called Explaining Computers helped change the world for the better. : )
Again, thanks for the video.
Great post, thanks for sharing.
My HDD from 2006 still works like a badass 💪
Excellent!
Maybe I should power up my old machine from 2005 to see how it's doing... :)
Been using it for 10 years before replacing it... with the same XP installation they've put on it in the shop!
i have hdd from 1996 , 512mb , he still work , but the data are from 2005
the 2004 WD10JPVX still works until now
Still works until now? 😂😂😂
You never miss your data until it's gone..... then you wish you had been organised :-)
Oh so true!
its ironic I havent lost a hard drive all these years but Im that type of person who never saves any data
Another great video! Thanks.
They can also fail if you move the drive when it's spinning. As with any spinning disk if you tilt it on an axis perpendicular to the spin it can distort the platter and cause a head crash. It also causes extreme stress on the bearings because, as you learned in school, anything in motion wants to keep going in that same direction which in this case is a straight line (around the disk). The faster the drive is spinning the more force it requires to tilt the drive. When you apply the force the platters resist that motion and place the load on the bearings and the platters themselves causing them to warp. (side note: server hard drives running at 10K or 15K RPM are two or three times as thick as consumer drive platters and are usually much smaller in diameter. This is why "large capacity" SCSI drives are usually of less capacity than "large capacity" consumer drives.) If you take a spinning hard drive (a bad one..) that is running at full speed try flipping it on its side quickly and you will understand. You will feel the resistance to that movement and you will actually hear the motor on the drive slow down due to the extreme load on the motor and bearings and may actually hear that terrible squeak of the head contacting the platter surface. This and dropping the drive while it's running are leading causes of mechanical failure due to head crashes. Many of the newer drives on laptops have a G-Force detector that can move the head out of the way (park) the instant it detects that the machine is in free-fall.You can test this also by turning on your laptop and tossing it (carefully) into the air and then catching it. If your drive has this feature you should hear the CLICK of the head parking itself and possibly the drive will shut down. Do at your own risk of course.
That scratching sound of the drive's head is better with headphones on :)
:)
That sounds pretty dope to me. If you found that piece of music I would give it a go.
Thx for the chuckle.
scratching*
I consider myself reasonably computer literate but I pretty much always learn something from your videos (including this one). Thank you for your efforts.
Your videos are very thorough and well structured. I love your channels. Thank you.
I've been using a HDD for storage purposes for over 10 years, and I use my computer everyday! Still working as new!
#1 cause of PC failures, for me, was motherboards, followed by optical drives, memory and finally, graphics cards and hard drives. Which is how I wound up with a heap of GPUs, HDDs, ODDs, PSUs and RAM; 7 (working) HDD made 2006-2010 total 830Gb, so their longevity is thwarted by improvements in capacity, sadly.
In that order? I found ram modules to be a pain, always lost some of them.
It's a blessing when SMART tells you that your hdd will gonna fail soon because you can backup your important data.
Many hard-drive sometime just "suddenly dead" even though SMART monitoring software like crystal disk info or HD Sentinel tell us that hard drive are fine or 100% healthy.
Should be common knowledge by now to not assume that anything labeled as "smart" is smart.
I have had a HDD and an SSD fail recently for unexpected reasons. One was spun up only for differential backup every couple days but had severe data deterioration and positioning issues at 10 years of age (masked because new writes from the backups didn't show any issue), and I had a Samsung SSD experience bit rot that the firmware is said to specifically avoid by refreshing old stored data.
There is two type of hard drive failure.
1.Seagate
2.Non Seagate
While I don't agree, your comment was hilarious!
@@Alfamoto8 two of my Seagate hdd died without any reason within a year of buying but another one which is named Toshiba is still up and running till now after two years of use...that's my story
@@syedkarimrayhan58 I have a 10TB Seagate HDD, Two 4TB and one 2TB and they are running fine, all our important data is in there.
I swear Seagate drives flake too quick.
I have bunch of Toshiba and it work for decades.
One thing that never seems to get mentioned is if your hard drive is starting up in a cold place. like if you have it on a concrete floor. it can seriously damage the hard drive if its starting up cold. that same goes for sudden temperature changes if you open a window in winter, that will kill your fans too.
It never gets mentioned because there are few to no applications where people are running HDDs in non-climate-controlled areas on a concrete floor, and even then, the floor has nothing to do with it. If you just mean a cool basement floor in winter, no, that is nowhere near cool enough to cause a problem if the room is otherwise a hospitable (to humans) temperature, and opening a window in winter does nothing to my fans, in fact I have ventilation fans directly exposed to outdoor freezing temperatures. They merely need a lubrication that stays effective in the temp range of startup as well as operational temp.
@@stinkycheese804 You are talking from your own experience it seems. My experience is that it does matter because i live in a place where temperature changes of 20c to 40c can happen if i want some fresh air by opening a window, clearly not something you've had to think about. Computers are used in all sorts of places, not just at home on a desk.
In the facebook data center where i work in Sweden, built here for the cold winters, we have to be very careful about temperature regulation as you may guess. May i recommend to you, look at googles study on hard drives so you can understand the lubrication is designed and proved to run best at 30 to 40c. temps below 10c can cause greater damage on start up, like anything mechanical, it will have an effect on its life span. The floor always matters in a cold room because heat raises... so naturally the coldest point is the floor..
@@ricky_pigeon you're right
Too cold is actually worse for than "too hot"
Solid info ! No-nonsense and straight to the point. Thank you.
And he is still hearting comments after 3 years?
:)
Very good basic introduction to hard drives. I would add that reliability greatly increased at least twenty years ago, though I've had a fair share of failures. I have had some drives running continuously for up to thirteen years. A 1GB system drive ran on a Linux server from 1995 to 2008 (with one CPU/MB upgrade during that time, and various Linux system upgrades). After sitting in storage it still ran again in 2014. Another server (a Sun 386i) ran from 1988 for about ten years. One of the characteristics of Unix-like systems is that they frequently rearrange data (a kind of defragmenting), so fade is not that great a problem. Different story, of course, for the external backup drives.
I still don't have any SSD's!
I find it amazing that you haven't talked about "disk metal extension" over time. Indeed for HDs that have been spinning 5-8 years (yes it's common on servers), the disk has a tendency of expanding. Thus if you have to turn off you disk long enough for it to cool down from its "nice and warm working temperature", it might shrink enough so the head becomes unaligned from it's previous format. Thus rendering the disk unreadable, and causing it to be unable to boot anymore. Rare but I've seen it twice in my life.
I have a Samsung 750GB SATAIII 3.5" drive that is now just over 10 years old been in 24/7 use for half that time or more. I still use it today, it runs every day, I save videos to watch later on it and use it for a temp drive when editing videos.
Samsung has always served me well with their drives, that's all I use and this is why.
Though this is the longest I've trusted a drive, I intend to use it until the end of it's life, still no bad sectors or heads and runs quite. It does use those fluid bearings I bet otherwise I would have been done long ago.
Good video thanks
A new Christopher Barnatt video, that makes my day good :)
I think it's safe to say that most people don't know about fragmentation. Also, I'm liking the pacing of your videos. Makes for easy watching or listening.
After working in IT between '89 and '07, HDDs were poor in the early years but after about 2000 get very reliable. The computers would cycle through the system in about 5 years so I didn't get to see the long term reliability.
That was a point I raised. The thing is, the HDDs will last long enough as you will more than likely replace it due to lack of space before you ever see one fail. In comparison, SSDs are virtually guaranteed to fail. Also, due to how SSDs operate, you cannot securely erase data from them you do not wish to be recovered like you can on a HDD.
@@NeilRoy They aren't, the TBWs are just "guidelines", the cells might often still writes but performance will takes a turn for the worst as the TBW leaves the "safe" zone. So as long as you avoid the cheapos (QLC and TLCs based drives) I think you can get good mileages if not better than HDDs mileages out of them.
At Microsoft, we used to call data fade "bit rot"
Good one!
And yet, replace the windows OS and the data on the 'rotten bits' of another drive reappears and is easily accessible, it's only Windows that rots; everything slowly stops working, yet windows reports that everything is fine. I have DOS and OS/2 computers from 1992 and the data never suffered 'bit rot'. My 10 y/o Linux system still works great. Only Windows seems to suffer from the OS slowly degrading over time.
These might be good for stationary uses, but anything that moves more than occasionally I'd rather go with an SSD. Sure they cost a lot, but not as much as data recovery!
One does not have to worry about hard drive failure too much when using external drive that most of time is not running or powered up. That said, because electronics age. A hard drive used to store important data should be replaced with a new one before it gets passed the 10 year mark.
Also super important data should be backed up to multiple devices, optical media and so on. Stored in right conditions and in a safe place. I find it incredible how little people back up data and just stor things in the cloud or just do not backup at all. The thing most people miss out on is that a backup archive or important data as mentioned in the video needs to be refreshed or maintained to avoid problems.
Great video :-D
Also, regarding heat sensors at about 11:00, Windows 10 software updates can and will sometimes remove temperature sensor based fan speed handling. Yes, you can find a machine automatically updated and stopped running fans based on temperature - and simply running fans at idle. So if you don't hear the fans they might actually not be running anymore. You can find a program to retake control of the fans/sensors and fix this - hopefully before anything's cooked.
even though i knew all this it wasn't boring to watch. u always go straight to the point and cover every subject perfectly. keep up the good work!
Thanks. :)
Excellent video sir. I didn't realise that memory fade would still occur even if the hard drive was regularly used or at least powered up. Although I now use SSDs in my computers, my NAS backups use hard drives, so I will check out the programs you suggest to refresh the data. Thanks for the tips. :-)
I don't know of many laptops that have HDDs these days (some of the real cheap ones do I guess) but I almost always recommend people spend the money and upgrade to an SSD (either buy a kit or have someone do the upgrade--although it's not very hard). This gives you the peace of mind of an SSD and more durability in case you drop your computer while it's running. Keep the old HDD as a backup just in case but I would recommend people use SSDs instead.
For cooling, the best option is to put a fan in front of the drive, that pulls cool air in the front and exhausts it out the back (towards the CPU, if your drives are located in front of the mainboard/CPU). This also helps with cooling of other components. If you can't do that, or there is no bracket or place to mount a fan, consider getting a 5.25" drive bay cooler (some also work as a 3.5" drive bay adapter) and use that. IF that doesn't work, then the drive rail with heatpipes could help. Most aftermarket cases should have a fan bracket in the front, but many OEM/store bought computers--particularly the cheap ones) don't usually have a fan located in front of the drives.
If you can find them, you may still be able to find a cooler with fans on the bottom that will blow air onto the circuitboard on the bottom, but I haven't seen these lately in computer stores (since most people are moving to SSDs or doing custom-builds with aftermarket cases that support HDD cooling via mounted fans in the case).
I always tell people that if a hard drive lasts a year, it'll probably last for about 4-6 more years.
I think that is a great rule -- and one I would agree with.
my main 500gb hard disk failed after almost 4 years . i lost incredibly valuable data . the issue was with the read head . either the servo motor or the whole head assembly was broken. i learned my lesson the hard way . please guys always backup your data
Reminded me of: _"There are two types of people: those who haven't done backups yet, and those who are, sadly, doing them already."_
:)
To serve as a confirmation bias to what you would have said 15 years ago, the only HDD I had fail was the one in my roughly 2 year old iMac. So your statement from 15 years ago still isn't very far off! :)
Thank you for the information regarding "data fade". I had no idea about this. It would totally suck to lose my backup data. Perhaps DVDs would do better for long time backup storage
Nah, optical discs are evil. Stay away from them.
Everything degrades over time. It’s best to store mission critical data on multiple hard disks or usb thumb drives.
RAID, my man
Thanks for another lovely clear fact-filled video! Perhaps it’s because I’m English but I find your explanations much clearer than many from over the pond.
This is the first time I heard of data fade on a hard disk. That's very worrying.
Tech Truth Do not worry about it too much, data rot will likely happen if it is very old or you put it in a magnetic environment.
Data fade is certainly real! Over time, the magnetic surface can lose it's magnetic charge, and this can result in "data fade".
Like he said, it happens for every type of magnetic storage. This obviously includes video tapes and K-7 audio tapes as well, that's also something to keep in mind.
I have drives that have been stored in my un insulated garage for 20 years these recently were checked if they still functioned. They still had data intact and no bad sectors reported after a scan and were able to boot linux and dos 6.22.
Higher density drives may be more subject to this than older drives. Scarily enough I still have Commodore 64 floppy disks from the 80's that are still readable.
Magnetic storage such as floppies and tape storage are more of a worry for data fade than hard drives as they are less shielded that hard drives.
I am more worried about my data stored on cheap optical media ie CD's from 20 years ago are becoming more unreliable depending on the storage environment and can be subject to delamination.
Tech Truth it's crap information
I have a hard drive that's 25 years Old when I got it it still had all the stuff and read with no problems
But the hardrive was only 200mb
Your vids are great, reminds me of 90s British tv when I was growing up. My hard drive is noisy as hell, only had it a year
Additionally, they dont make them like they used too. (As they say). I work in a DC, and we swap out maybe 30 to 40 dead drives every single week. The new servers with 750GB 15k drives tend to fail after about 1.5yrs. However for specific reasons we have some really old compaq ml370 servers with 9GB 15k drives that are still working fine to this day. Some 17yrs later !.
Perhaps that's because the newer drives handle much more data than the old ones and are thus more likely to fail?
Same here, same exact disk drive and Compaq server and it all still works. Removed from service simply because obsolete.
I cannot know everything, but I do try to keep abreast of things that I may not be so attentive to. I learned from this video quite a lot.
my favorite tech PC tech guy on youtube :D seriously linus is just a joke :))
I don't like Linus' "american reality showmanship". Everytime I watch a video from that guy I get upset and stressed.
Chris' videos causes me the opposite effect though. I feel relaxed and ending up with more knowledge that I begin with.
Both have a lot of knowledge. But Linus can be irritating sometimes.
For me, this was one of your most informative videos yet. As a bench tech I'm always challenged when explaining to a casual user why a seemingly good drive should be replaced based on SMART data, you've given me some great information to pass on in simple terms. I'd be curious as to your opinion on the accuracy of some of these utilities that go as far as predicting the approximate number of days until failure. I use hard drive sentinel and on several occasions I have encountered drives that allegedly have days left. For fun I've run these drives in test machines and of course they don't fail on cue even months after they've been reported to expire.
Thanks for this kind feedback. I don't think any utility can predict the number of days to failure! Rather, the best that can be reported (or taken from such days-to-go predictions) is that a drive has a reached the point where it has a high probability of failure, and should hence be retired from critical use.
RUclips really gets me paranoid about losing data today! Know my job tomorrow - backuping my stuff! 😂
Backups are always good. :)
Thank you! Data Fade prevention is something I had not thought about!
Just downloading DriveDX Trial version as a Mac alternative and will spend the day going through my drives and back-ups - thanks!
SMART was originally intended to catch predictable failures but in actual implementation it rarely catches anything until it's too late. This was apparently due to concern that such predictive information would likely increase in-warranty repair attempts and could provide necessary legal evidence for class action lawsuits. Instead of reporting failures before your data is lost the drive plays musical chairs with failed sectors until the unit has passed well beyond the warranty period and the defects begin to exceed the drive's ability to hide them. At which point it's time to buy yet another drive from Western Digital (which now includes Hitachi and Seagate).
When noodling around in the linux man pages, I saw some peculiar, forgotten, or ignored stuff. At one time, in the old days, the way to take care of bad sectors, which normally increased, was in the file system. It wasn't internal to the HD. I believe you can still have failing sectors marked for non-use, if you care to.
Another thing: At the factory, before you get the HD, they map out bad spots, because all platters have them.
Thanks for sharing. I have found it essential to keep old computers working and your little study will help in that respect.
I learned alot with this video! This is my favorite channel. :)
I bought a 1Tb Samsung HDD in 2011, used it for 4200 hours on my desktop, in 2013 I put it on a dvr, and in 2015 according to HD sentinel the performance was at 100% and the health at 0%, I didn't replaced it because I want to see when it will finally die, yes It is still alive and spinning 24/7, with more than 40000 hours, the model is HD103SJ, manufactured in 07/2010. By the way, the temperature inside that dvr is above 50°c, hard drives are great, I had an SSD and went back to an HDD. I like the channel.
Que videos tan chingones! Gracias.
Voy a estudiar ingeniería en sistemas y robótica espacial.
Thorough and useful vid thank you. Wish I had this info a decade ago; could have saved a few HDDs.
Iv ready somewhere that overdoing HDD cooling (below ambient) also reduces it life expectancy.
(maybe due to condensation inside?)
Iv been using HDsentinel for many years now. Iv also used reports from it to get a few disks under warranty replaced.
Still i have lost 2-3 HDDs due to failures so far .
Another very important cause of failure i have noticed is abrupt switch off of HDD due to any reason like system crash/loose contacts (ext. HDD). When this happens too often bad sectors accumulate over the disk and that very much reduces disk life.
Apart from this, iv noticed, when the HDD hardware is apparently defect free the main culprit is temperature (esp. in laptops).
i had my computer for 10 years it's still working
Yeah I had an old HP desktop from the 2000's and it lived for about 10 years.
@@techgamer1597 its dead now lol my PC won't turn on anymore mostly because of porn
@@fightfannerd2078 thats why you dont use windows for porn, use a mobile device instead :)
I have a 17yo eMac that still runs fine, but I guess buying vintage computers on eBay is pretty iffy.
My computer is now over 10 years old (bought in August of 2010), but initially I only had a prebuild that would fail about 4.4 years later. Then I used a laptop for ~2.8 years, and after that period I finally built my own PC using working parts of my prebuild (CPU, RAM, mainboard, hard drive, DVD/CD drive) and bought the other parts online (case, power supply) since the ones of my prebuild were too small / weak for what I had planned.
As for the graphics card, I was lucky and got a Zotac GTX 470 from a friend of my sister which I used for ~1 year although for some reason after a summer vacation (I had it for ~8 months already) it suddenly "died". However, I could still use it without drivers but it stuttered _a lot_ and I think sometimes it would even have visual artifacts on the screen. Well, needless to say I could barely play any games or even just watch videos / browse the web smoothly with that "zombie" of a card...
Fast forward a few months in the autumn of 2018, I got my new graphics card at last! It is an MSI (AMD) Radeon RX 570 8 GB Armor OC - which I still use today. :)
So this is the current state of affairs. Sorry btw, this turned out to be an essay lol
Wow thanks so much for this, I came here out of curiosity (I did know that HDDs lasted a while if cared for right) but I DID NOT know of "data fade" at all. Now I do and will be refreshing my backup drives at external locations 'refreshed" every year or so. It's mostly photos, but honestly they mean a lot so I want to keep them safe and backed up.
havent seen ur vids in a while!
EDIT: awesome vids BTW! keep up the good work!
Thanks. I post every week, so click that "bell" icon thing to be alerted every Sunday! :)
Having trouble booting up , ran a test and it says everything is ok except for the fan . Anyway won't let me start , boot up log on ,,,nothing . How can i recover the files , pictures , docs from my old laptop to my new laptop ???
Still here with my 80GB hard drive on my ThinkPad X200S.
Though my Grandad still has his hard drive from the early 80s going in his really old laptop. It's quite impressive.
I never lost anything to a HDD failure....said no one ever.
Indeed.
For me, not YET. That is why I frequently backup my data.
EJ Tech so do I but I lost some data when I was young, did you never lose any of your data to an HDD failure ever?
that's great, hopefully you'll never have to suffer data loss. :)
I only ever lost data to ONE failure of a HDD...but that was me being stupid and over voltageing it :( I Will most def get the disk fresh prog though since I pretty much do just shove my HDD into a cupboard. Lol
I'd never heard of data fade on a hard disk. Thank you, sir. I have been using a 2TB hard disk to hold media files, as well as a semi-regular backup image of my OS drive for a year now and will get refreshing software just in case, since a lot of the media files are permanent and probably never move.
I remember how we used to park heads.
P for park. ;)
C>PARK_ (Imagine the underscore as a flashing cursor, lol)
Now WD green parks the head for you, my had over 600000 last time I checked.
But yes, I herd that power failure was a dead sentenced to a spinning drive, at that time.
I don't know how common they are, but there are drives that have an acceleration sensor in them. When a drive falls it will detect this as a few milliseconds weightlessness and park the heads before the drive hits the floor.
Of particular interest to me was the concept of data fade, and refreshing your data to prevent that. I had never heard of that. Thanks!
I'm also about to run the Data Lifeguard Diagnostic (after I close this browser window) :-)
My hard drive is still running great after 7 years!
Excellent! here's to 7 years more!
Mine still going strong after 7.5 years! 😎
I tend to dismiss "data fade", at least in the case of much older HDDs. For example, there are many videos on RUclips of people recovering very old computers, from the early 1980's, with hard drives ranging from 5 to 80 megabytes. These drives usually have sat, untouched, for many years. In a fair number of cases, the data is still accessible. In some, stiction, motor failure, or even a combination of both has made the drive useless, but that is by no means a guarantee of failure.
Having said that, I partially agree with your recommendation to periodically refresh the data, though I would just regularly move the data to more modern capacity drives or formats.
13:06 "Back in the late '90s and the naughtys" Did anyone else chuckle at that? I mean, I know he probably meant the 2000s but still.
The "naughties" is a common phrase for referrng to 2000 to 2009! :)
Yes, I caught it immediately and chuckled, though I understood it too, since naught/nought is a British term for zero.
Excellent Teacher with amazing teaching style and methodology.
I used to have a cat that kept knocking my external hard drives off my desk, I gave the cat to me aunt, and now I have a dog.
Sounds like user error. Dogs are generally much more destructive than cats. I know: I have both.
u seen that meme to XD
I almost have the same HDD that is shown in the video that's just sitting in my drawer. It's the WD Caviar WD200, 20GB. I don't know how old is it but I know I connected it to an older computer something like 4 years ago and it worked, booted up Windows XP. Now I'm thinking to get a SATA-IDE adapter and connecting it to my PC to see does it still work.
Great video, explained right to the point. I should really have more attention to my current HDD that's about 7 years old and possibly get a new one to be on the safe side. :)
HDDs in my experience are tough resilient beasts.
I have a bunch of really old drives from the 1990s ranging from 10 to 80 GB that I have submitted to constant full erase/ write cycles performing back ups on a close to daily basis for 9 years now.
Intending to test their durability I even have had them operate with poor cooling becoming really hot, reaching almost 70c and operating like that for hours.
half of those drives were reported as imminent failure by monitoring software with less than 12 working hours left on them when I first got them.
other than re magnetizing a handful of sectors via software on a couple drives
None of them has given me
any trouble and all have worked way longer than the predicted 12 remaining hours, even the "crappy" IBM travelstar drives.
really? thats good to hear but i will still back up my stuff every year just to be safe
i'd recommend buying an m.2 nvme vnand ssd because they're fast and expensive
this video style made me thing this video was made 15 years ago but i was pretty amazed to see its only about 6 years old
Datafade is not a joke. I am mega paranoid about it, full backup refresh every 6 months. Just be sure to refresh backup disks one at time, use a surrogate disk as a temporary backup to your backup in case of failure during the refresh. If you really want to retain your data, have a distro stick to boot from during file copy operations with your backups. Dont use windows, Microsoft filed patents to compare your files to known hashes they are paid to look for and auto-delete them. Its still in the development stage, but it will eventually make it into the update pipeline, and I suspect they will not tell you. Plug your backup with all your movies and music into windows 10... and.. its gone! That's what they hope to roll out. Dont play games with your files, do data management and wrangling in linux.
Lol, the new evil MS mission: to destroy everyone's backups.
it's only paranoia. Rewriting all the data every 6 months is straight up insane, even every 5 years makes no sense. In fact, it makes no sense to do that ever, since the drive can fail anyway. All you need to do is to check that the data is readable, nothing more.
And thinking that Windows will read your every file, really? That would absolutely never happen.
What's this patent you mention? I wasn't able to find anything related to hashing the contents of files.
I have a Toshiba P300 4TB and I have no problem with it and I'm really happy. I hope it will serve me a long times. I also have a Toshiba m.2 ssd.
Excellent video
Good view to talk about reasons hard drives fail. However, there are no answers on life expectancy of popular hard drives.
Now imagine this guy hyped AF like linus haha
If an HD starts ticking make immediately a backup, don't put your pc off!!! Or it's dead, In my 40 years been busy with computers it happened a lot. Although I had a few where I lost data, I always use enterprise HD's, but they also fail after an amount of time. Thanks for the clear tutorial.
"click of death' is a 2020-thing?!? ;-D
Those rubber mounts for internal harddrives are to limit the noise produced by the drive vibrations, not to protect the drive itself. The absorption of these mounts is so limited that they won't help against falling or shocking.
Wished you were my high school teacher.....such a detailed video...awesome.
Thanks.
had a head crash on friday... not fun stuff I can assure you that and recovery services are quoting me 12 times the price of a replacement drive...
Sorry to hear this. Recovery from damaged storage hardware is always very expensive.
If you fancy the DIY level of cooling you can use a aluminium/stainless steel kitchen pot scrubby pad and (if you use two drives on top of each other) piece of plastic to shield the drives PCB from shorting out. The scrubby will take away the heat from the top of the drive which then can be carried away by means of a fan or series of fans.
please make another video about #-lithium ion battery life expectancy
Now that is a great idea, and a subject I know a lot about! I will now certainly put such a video in my schedule. :)
ExplainingComputers I second this!
ExplainingComputers can you add NiMH to it too?
i have a short answer for you. 1000-3000 charge cycles depending on the compound used as a cathode.
i like that to cause my kindle fire hdx 3 yrs old and battery still good
Years ago, back before many of you were born and before machines like the VIC-20, TRS-80 or the Commodore PET, the relatively non existent computer industry (there were big firms like IBM and Honeywell making huge mainframes but nothing for the home user) was engaged in an informal debate as to how to cheaply implement computer technology into homes, hospitals, elementary schools and small business. One of the initial topics was memory. Industry used large tape drives or punch cards to store programs and data. The fledgling home computer industry didn't want to go this route. Their idea was to store as much data as possible electronically without cards, tape or a very primitive optic drive. The problem was it couldn't be done. The cost was too prohibitive and the technology to cheaply manufacture high density solid state electronics was just not there. Technologies after the development of the transistor such the Op-AMP and the four NAND Gate chip were HUGE. So the industry had to turn to cassette tape drives, floppy disks and eventually CD/DVD burners and HDD hard drives (in early drives those disks were made from glass and weighed a "relative" ton causing many spinning motors to burn out after less than a year of use) . The SSD drive was where the consumer computer industry wanted to start and it took almost 50 years for it to get there. Today the HDD is on it's last legs. The detour is finally over. Already gone are punch cards, tape and floppy drives and almost gone is the DVD burner. The only reason the HDD is still readily sold is due to inventories and third world manufacturing costs. Once the inventories are gone so will be the HDD. For all my clients I am switching them over to SSD. Any HDD implementation is done externally. I am much older than the presenter of this video and I can say it has been a fascinating ride these past 63 years. When I started working in consumer home electronics a home computer was just an "Orwellian." concept. The Vacuum tube would still be king until 1970 and an ipod was a transistor radio with a crystal earphone. I will not live to see it but I believe some form of implantable computer technology (either biologically or electronically) will be dawning soon in the next 20 years
Any HGST lovers here..?
Al Ameen p s mines been working since 2009 never failed
It's all about probability. Mine 2tb seagate with reputation of dying in months survived four years to first bad sectors and five years before it failed with ticking sound. You can't make any statement about reliability even if you have ten drives, cause you may be lucky or unlucky.
82GB HGST here!, since 2007 :)
Lucky
My Asus Republic of Gamers laptop came with a 2TB HGST drive as well as an m.2 SSD, but I kept the drive because it's super fast and very reliable. I added a 2nd 2TB Samsung drive as well, but it just hasn't got the same performance.
My WD drives usually go about 40,000 hours on them before bad clusters turn up or they start to get noisier or sound unusual. They usually still work fine at that point, but that's when I know it's time to replace them. At that many hours, the slightest sign of failure means there's a good chance they may fail soon & possibly dramatically. Loud & excessive clicking means failure is imminent.
Yeah. Planned Obsolescence. It's been a 'Thing' since the mid-late eighties unfortunately, and perhaps even earlier in some places..
Hard drive failure isn't planned obsolescence unless the manufacturer designs the drive to fail. In 95% of cases it's because of wear over time; the laws of physics and thermodynamics still apply, and that's something that can't be helped.
I always do a surface test (any software will do - I use Victoria, for example) of an HDD when I buy one - even if it's brand new.
Because better safe than sorry.
And one more interesting fact. The less you turn your HDD off/on - the healthier it will be (if it doesn't doesn't overheat from running 24/7, ofc - so use proper airflow). It's because of the expanding and shrinking due to it getting hotter then colder. Those micro-expansions will damage the actual hard disc inside the drive. And the motor (as any other motor out there) is most stressed when it spins up (starts up), so it also help to preserve it.
That is an another reason I set the HDD spin-down timeout setting to 0 (disabling it) after Windows installation.
Thanks for the nice content
Bit of a sigh to open by saying HDDs fail for many reasons but SSDs only when reaching max specced writes. You even include manufacturing defects for the HDDs but don't consider them for SSDs.
Also, rubber mounts make very little difference. They are not shock absorbers, so they can only absorb high frequency vibration, and even there you'd need more elaborate constructions to isolate the drive from the case, and usually you have to do that to shield the case from the drive's vibrations more than the other way round, because most harddrives vibrate noticeably.
As for harddisk coolers, it is much better to pass on that (especially one with little surface area as the one shown) and instead put the drive in a fan air flow path. That alone should suffice for any situation, even just a silent fan. Very mild direct air flow is multiples better than none at all and still much better than passive air flow from case underpressure currents or such.
But SSDs are fairly likely to fail simply from component failures, which happen unexpectedly and can cause data corruption from mild to severe, or render the drive completely unrecoverable.
While exceeding the write capacity is something you should be able anticipate, since you can easily check how many writes you've officially got left.
And having an SSD stop working due to exceeding write capacity is not scary at all.
All the data can easily be retrieved without problems.
So one should only care about component failures, not at all about exceeding the write capacity.
Most things are designed to break!
like anything apple and anything with secureboot in mind like ios/android devices and hopefully we can avoid the storm is windows 10 s devices
Like people ;)
anything that stores data should be designed to last as long as possible (what if you are working on some groundbreaking research and you lose EVERYTHING because your drive was designed to fail)
it should be designed to last as long as possible but that is not the mind set, its to have more of an guarantee of an "upgrade" cycle so you buy more drives more regularly, and freak people out and make them use raid 1 instead of 0 for their personal desktop that does have a backup
Andrew - possibly but not universally true and everything is subject to inevitable 'wear & tear'; but a wasted comment I'm afraid as the critical aspects are WHEN and UNDER WHAT CONDITIONS, (also, I guess, what can be done to delay the inevitable?)
PVR is 12 years old which is a external HD sun into receiver. Been on constantly and I'm amazed it's still.working.
That is indeed amazing!
Damn i wish i knew that hdds do eventually die. I had a pc that fried its motherboard so i stored the hdd. After a few years i got a new pc but my old hdd is not working :(
I also had an external hdd with thousands of pictures and it failed too very recently. So if you have an hdd with your valuable data in it, do yourself a favor and make a backup.
For testing a drive, I use the “badblocks” utility that comes standard with Linux. It can do read tests and write tests, and also keep repeating the tests as long as new errors keep showing up.
And it’s not vendor-specific--it works with *all* drives.
SMART is a waste of time.
My Toshiba hdd fully dead just in 3 years.
RIP - Rest In Platters
This story is a great forum about hard disk drives. I hope to see a show about flash memory drives. I enjoy your program very much.