Why you Shouldn't Low Level Format Your Hard Drive | Nostalgia Nerd

[@measl]

*Nerd is exactly right on all counts here. In the MFM, ESDI, and early RLL days, it was not uncommon for people to hide data by formatting drives in unusual ways, and then writing and reading their "secret data" to the **_"extended sector"_** areas. Yes, you read that right: the **_"extended sector"_** areas ! The scheme goes a little like this: you set up the sectors to be twice as big as expected (just an example, you could use any multiplier), and then write the sector headers to only show the regular sector size (usually 512 bytes). So if you have sectors claiming they end at 512 bytes, you have another multiple of 500 or so bytes there which can be reached with custom drivers (you don't get all of the secret space because there is a little overhead required which eats a few bytes of the secret data). Your drive works as expected, but it just appears to be smaller that the label says: this was not completely unexpected either, since the early drives could have any number of "bad" sectors, and an examiner would assume that bad sectors were the cause of smaller capacity. Note that this is not the same as the earliest attempt to hide data by claiming good sectors were bad, and then using them anyway (again, with a custom driver which you had to write yourself) - this idea was old enough to be easily discovered by any competent systems programmer.* *Finally, there were the true secret data "pros" - they would format a drive normally, but specify a small number of extra "timing bytes" to separate each sector, effectively adding a [small] fixed number of extra bytes **_in between each sector_** ! This was a difficult system to use, requiring a great deal of technical knowledge, as well as a lot of assembly level programming skill, but it did provide a small secret "file system". The intelligence agencies and big corporate users were likely the only ones using the inter-sector system, and if the drive were ever written to using a driver other than the one written for that particular controller and drive pair, the secret file system would be instantly destroyed - this was a kind of bonus in functionality. If the secrets were dangerous enough to justify using the inter-sector system, then the data was probably worth destroying if it fell into the wrong hands, and this system pretty much guaranteed that would happen. I worked at a place that used the inter-sector system on their [XENIX] server for a few years in the early through middle 1980's, which is how I even know of these crazy things.* *Thanks for a fun look down memory, er, "disk memory" lane, Nerd!*

[@Sypaka]

If that wasn't annoying already, you could set an entire partition to hidden. it completely vanished as it subtracted it's size from the information table and I don't mean just removing it's drive letter. No one will notice one of those on modern HDDs. 200 MB missing? who cares, 4TB, bitch! Back then you needed special tools to unlock them. I think it was called EISA or so..

[@chisel4164]

*WOWZAS*

[@chrismc410]

MFM and RLL...certainly shows your age. Likely 40ish or close to it, minimum.

[@TheBashar327]

@@squishyu , just ignore the *special* children's ticks.

[@TheBashar327]

@measl, that was great information.

[@factsverse9957]

Then: $3398 for 10 MB! Now: $0.03398/GB!

[@LoneTaurus82]

How times have changed. How anyone thought a 10MB drive was worth that back then is beyond me. That would equate to 2 or 3 high bit rate mp3 songs today.

[@dylanharding5720]

@@LoneTaurus82 because they actually were worth that. For example, in a few years the 1080ti will be cheap. Really cheap.

[@chrismichaelgardner]

@@LoneTaurus82 People were storing just text. 10 MB is a lot of text.

[@timramich]

@@dylanharding5720 The GTX 10 series came out almost 3 years ago, and those few higher end models are now more expensive than back then.

[@coomcake]

This won't age well

[@aitchpea6011]

Used to work at IBM, building hard drives. The "Low level" format you're talking about was known in-house as Servo Track Writing or STW for short. It was performed on the drives in a cleanroom before the top covers, seals and logic boards were attached. The STW process required a large machine with a big, heavy, shock-mounted granite slab on which all the moving parts were mounted. IIRC the process took about 3 minutes for a 3.5-inch, 4-head, 2-platter disk which could hold, I think, 15Gb. The STW machines were delicate and went wrong at least once per 12-hour shift.

[@ToTheGAMES]

Don't low level format? Nice try, NSA.

[@lewisfilby2394]

im sure he's joking

[@gregorykhvatsky7668]

I think DBAN is a much better data destruction solution

[@chrismc410]

Gregory Khvatsky DBAN, Degauss and physically Destroy. That's the proper way to do it. One place I worked called it The Three D's

[@TheSliderW]

Use a tool or script to replace "empty" space with random junk data. You can even do that as a planned task. If they're banging on the door, then unplug your drive, smash it with your hammer and throw it in the microwave oven for good measure. Be sure your SSD's don't have any metal shielding them, smash the chips individually. If you're hardcore and they got you by surprise, then you already sacrificed a microwave oven to mod your PC so that the magnetron is pointing directly to your drives and you just have to press a safety switch or ask Alexa to "turn on the fireworks".

[@gregorykhvatsky7668]

Then dump what remains into a black hole and wait till it evaporates through Hawking radiation. This will ensure the data is destroyed properly lol.

[@KuraIthys]

I remember having a bad drive and attempting to find out about the arcane art of data recovery. There are many, many things they refuse to tell you, but one thing I did discover is that the drive I had a problem with had a secondary data connector hidden among the jumper pins. This connector allowed you to do such things as change the drive parameters and other firmware stuff. Normally messing with this is a REALLY bad idea. Because these parameters determine the exact alignment and logic the drive uses to read data - this is set in the factory, but the drive itself sets various parameters over time all by itself as the drive ages and content changes or to adapt to various mechanical quirks. The result is, no two drives have quite the same set of parameters. So what happens if you need to do data recovery, and the controller board or worse, the head mechanism is damaged? Well, one option is to replace these with parts from an identical drive. All well and good, but all these hidden parameters that are unique to each drive? Once you've done this physical replacement, you can bet nearly all of those parameters are completely wrong. So what to do? Well, change them one by one with this extra control interface, until you get a set of parameters that works correctly again. Tedious, frustrating, arduous work, but that's what data recovery turns out to be like. (that, and you need to do it in a clean room because if you open the drive, dust will kill it in a manner of hours at most.) This interface also allows you to read low level diagnostic information. but of course, this protocol is unique to each drive model, requires specialist equipment to access and is not designed in any way for use by regular people. (it's factory diagnostic and calibration data, ultimately...)

[@theLuigiFan0007Productions]

The same thing also works for one drive with a dead logic board. But, it only works for drives with a close enough revision. Get our the hot air rework tool, desolder the configuration flash from the dead board. Swap it onto the board of a functional drive of the same type, then put the functional PCB onto the drive that had the dead PCB. Chances are it's very likely to work if the configuration chip isn't dead. Not guaranteed, but worth a shot.

[@Ccoolty]

Why would you need a clean room if you're only replacing the logic board?

[@askjacob]

You make the debug serial port sound like some nefarious secret. It is not that secret or scary

[@KuraIthys]

Ccoolty - You wouldn't. But in many cases, if the drive has had, say, a power surge, or blown something on the circuit board, the voice coils in the head assembly are also fried. Replacing those requires removing the cover from the drive. There are also several other last resort data recovery techniques that require it. askjacob - You obviously have never spent much time trying to get information about how this stuff works from data recovery experts. Who knows, maybe they're more open about it now, but 5-10 years ago getting any useful information about... Pretty much ANYTHING to do with drive mechanisms or low level logic was like getting blood out of a stone. These people guard that information like a hawk. Or they did, in any event. Granted, one a decade or so ago did explain in detail how the debug port on many seagate drives worked. But trust me, 10 years ago, it WAS a nefarious secret. As in, nobody would tell you anything useful. At best you'd get vague hints about various bits and pieces. Anything to do with data recovery techniques constituted a jealously guarded secret. If things have changed at all, well, good. But that WAS the situation when I was trying to fix one of these drives. You'd have an easier time getting the internal microcode for a processor for intel than you would getting any but the broadest bits of information about hard drive behaviour and functionality.

[@measl]

@@KuraIthys Of course! We made a LOT of money doing recovery, why give away the family jewels?

[@berczigabor]

The Master Boot Record (MBR) has nothing to do with disk geometry or low-level formatting. It's like any other sector on disk that's only accessible after the disk has been low-level formatted and it's geometry established, and contains the partition table data. However, contrary to what the video states, it does not contain the disk geometry data. At the time the latter was important, this information was printed on the disk drives themselves, and had to be entered into the BIOS, which stored them in CMOS (and again, not on the disk, which would have made no sense, because the geometry was important to establish the disk layout needed to read it in the first place).

[@AltoidJTP]

Correct. This video, and the comments, are really bringing me back to my days with the original PCs. I really do miss those days.. SO much fun tinkering with computers at that level.

[@peterlamont647]

I actually found it easier...lol. At least you _know_ how they worked back then. Not with 30 trillion layers of obfuscated abstraction on top. i mean I have a book on the windows driver foundation. It's two volumes! Topping out at over 600 pages each...RIDICULOUS! If anything in that massive cloud of swarming bees gets out of order, it just gives you a coy message saying there was a problem configuring your hardware. The old days: Bios, go there and get the stuff. Got it! Let's check the memory out. Okay let's find an operating system... Found it! All done. Want a serial port? I can point at it. It has an address! It's at 0x03F8h no problem! Want MIDI? Great, 0x0388h Today: Bios, go there and get stuff. Got it! Let's check the memory out. Okay let's find an operating system... Found it! All done. Want a serial port?? NERD! F#CK OFF... NO SERIAL FOR YOU. IF YOU ASK NICELY ILL TELL THE KERNEL YOU WANT TO DO SOMETHING BESIDES USE MS WORD. LOSER. I CAN'T BELIEVE YOU WANT TO _USE_ YOUR COMPUTER...WHO DOES THAT?!

[@night_fiend6]

Peter Lamont there is your problem, Windows.

[@NelsonBigGunP200Fan]

pretty sure at the factory the drive is prepared with machines, or a special computer, so you wouldnt need to llf a drive when u get it inside your pc

[@phillipsusi1791]

No, the MBR certainly does contain the disk geometry ( CHS ). This did not actually set the geometry of course, but was there to tell the system what the geometry IS. Without knowing the correct geometry, you could still read the MBR because it is always at sector 0, head 0, cylinder 0. Having read the number of sectors and heads, you could then correctly address sectors not on head 0, clylinder 0. Hrm.. actually now that I check for a reference, it appears it was in the Bios Parameter Block of the FAT filesystem, which was only the first sector on floppy disks, whereas hard disks have the MBR. I could have sworn that the MBR also had the BPB but apparently not. Or maybe they did, but only DOS considered it to be there, whereas the "standard" format called those bytes just part of the boot code. Probably that one.

[@PeterBrockie]

Fun fact: Back in the early days of the Macintosh (probably PCs at the time, too), it wasn't uncommon to low level format the SCSI drive for your Mac Plus or other slow-ass computer of the time to NOT have the sectors sequential. You would have an arrangement of them where it would look sometime like: - - - - and so on. This was known as a 3:1 interleave. This was because the computer's SCSI connector was so slow at reading the drive that it wouldn't be able to keep up with reading each sector so quickly. The staggered sectors gave just enough time for the CPU to process the first sector before the second one rolled around - otherwise it would always miss the second sector and have to wait for it to spin around again causing a massive performance hit. As hard drives started to include on-board buffer memory this became a non-issue as it was easy to trickle out the sector data from the cache.

[@OneEyedJack1970]

Yeah, drive interleaving. That's one of the things Spinrite was supposed to optimize for you.

[@PanduPoluan]

Ahhh interleaving! I remember formatting a hard disk several times with different interleaving values to find the best. Also, inter-track offset! You want the 'next' sector in the 'next' track to be just about to enter the head's field after the head shifted to the next track and prepares itself for read. So much arcane magic...

[@n2n8sda]

This title is a bit misleading? Shouldn't it be "Why you can't low level format a modern HDD" as you even go on to explain that modern drives can't do it anymore due to the inability of the write-heads. Only very old drives that were designed to be low-level formatted or some very early modern ones will actually low level format to the point they will attempt to re-write physical sectors.... I can't remember what it was anymore but perhaps IBM DOS but seeing that bad-sector sticker on the drive brought back a memory of when formattting the disc the software would ask you to input all the bad sectors marked on the drive before you started, that was also a pain!

[@peterlamont647]

This was before scan disk tools were in wide use that you had to manually enter junk like that. You can low level format a modern drive. If not, how does it flag all bad sectors after the format? This is as true today as it was back then. The major difference is the sector size. Making allocation blocks larger allowed for these scans and to be faster. Try running a detailed scan of a floppy under dos! Agonizing and slow, but it flagged all bad sectors and you had a known good disk to write on with confidence! The title is misleading, and the content is straight up wrong! lol.

[@SkyGodKing]

He explained this in the video. Performing a "raw low level format" would break the drive, so the firmware doesn't allow you to do that. The firmware shows you what it wants you to see rather than the raw data. This video is pretty much pointless and misleading, so don't bother thinking about it in much detail. The only point of interest would be the fact you don't have full read/write access of the platter data means that even if you use DBAN or any other tool to "low level format" your drive there may still be sectors that are completely untouched by the format, etc. e.g. if a drive marks a sector as bad, then you can't overwrite or clear that sector. Some drives reserve space that may be used for other purposes, catch, or for remapping bad sectors, etc., which is similarly inaccessible by the system

[@bryanlatimer-davies1222]

I agree if it had a bad track table you could and should low level it IDE and later not quite the same process at all

[@n2n8sda]

You raise an interesting point there.. I assumed (perhaps incorrectly) that during a full format of a modern hard-drive bad sectors are re-tested. This is all very hypothetical from here and nothing the average user should ever worry about but I wonder if sectors marked bad are not re-written then if any previously stored data on said sector(s) could be recovered with approriate hardware / software if the sector wasnt too badly damaged.

[@mapesdhs597]

Some UNIX systems (Linux too?) can low-level format a drive (the fx command on SGIs can do it), though doing so may spit out a huge warning stating that such an action is almost certainly a very bad idea, takes a very long time to run, risks ruining the drive and possibly the end of the world. :D Still, it can be a useful function sometimes, eg. a lot of Enterprise SCSI/FC drives are formatted using 520 bytes/sector, but 512 is better for end users (standard OS/tools may not understand drives that use 520, they're often intended for dedicated enclosures/controllers/etc.) Thus, being able to do a low level format of such drives back to 512 is very handy indeed, but as I say this applies more to SCSI/FC (perhaps SAS too). Friend told me he bought a 4Gbit FC enclosure recently containing 14x 300GB 15K FC disks for less than $150 total, low level format to put the the sector size back to 512, voila, large array ideal for use with IFFFS (uncompressed dual-stream HD/2K).

[@RetroBerner]

Thanks, could have used this video 20 years ago.. LOL

[@RetroBerner]

Abbrahan beta max or bust

[@ajddavid452]

dvd players were upwards of $1000 in 1997 better give him a vhs

[@rogerwilco2]

We used books 20 years ago. I had a 800 page one called "Upgrading and Repairing PCs" it was awesome as it described all details of every model of IBM, Compaq and HP up to the early 386's. I think it was by Scott Mueller.

[@mikehikes710]

Steve Martino )

[@jedits1988]

In 1997? My first drive was an IDE (37 Mb) back in 1992 and I never had to low level format that either.

[@garyharrison4915]

Hard drivers are still unbelievable technology.

[@wingracer1614]

They really are. It amazes me that they can make a mechanical device with that level of precision and endurance for $60.

[@cdreid99999]

Actually they were and are incredibly simple technology. If you actually look up how they work you'll see why. Just as crt TV's were

[@MephistoDerPudel]

@NPC 363-7654-80521-Beta So, in the 90's they had TFlops of Performance in their Desktop-Systems?

[@elixier33]

​@@cdreid99999I'd love to see you make one then.

[@pcuser80]

In the dos era for mfm drive you must low level format A:\debug G=C800:5 Fil in defect list and go A:\FDISK reboot A:\FORMAT C: /s C:\

[@n2n8sda]

AH!! I posted a comment above asking what it was where you used to have to put the defect list in off the label on the drive... i couldnt remember if it was in dos or not.. It's been a while since i've had to do that (machine ran IBM DOS) so I thought it might have been the format command as i don't remember doing it on later machines running MS-DOS (probably rll hdds) but must have been the drives low level format utility!

[@pcuser80]

The low level format routine makes the sector unreadable, at high level format dos puts a marker (bad sector) in the fat table.

[@mbunds]

pcuser80 - The Seagate ST225 (20 Megabyte) and ST238 (RLL 30 Megabyte) used debug entry g=c800:ccc. I actually have a couple functional units, one in a classic IBM PC-XT equipped with a "turbo" mod that over clocks the 8088 from 4.77 MHz to a scorching 5.2 MHz, and is also equipped with a massive Megabyte of RAM using all four banks of 41256 RAM chips (that's 32 sixteen-pin DIPS, each of which had to be inserted into the motherboard sockets, one at a time). A person can type nearly as fast as these machines could scroll text. Oh yeah; it is also equipped with the venerable Hurcules CGA graphics card!

[@0311Mushroom]

+Mark Bunds the DEBUG command was determined by the controller card, not the drive. MFM/RLL drives were very stupid, you had to access the controller card to do low level formats. G=C800:1 was the usual command, but some used others. Seagate was the most common one, and most copied their commands.

[@pcuser80]

My card was C800:5 I remember a old trick to use a MFM drive on a RLL controller. 20 MB > 30MB (free 10 MB Extra) Not very reliable but it works, choose a MFM drive with zero defects.(rare)

[@JimLeonard]

To clarify, you absolutely must LLF MFM and RLL drives, especially when introducing them to a new controller. It is only IDE drives you shouldn't mess with.

[@AltoidJTP]

I ruined a few IDE drives that way. Though in my defense, they were already being problematic, and my old school way of thinking figured a LL format was the logical next step in an attempt to mitigate the errors. Didn't take long for me to actually do some research and discover why that was a bad idea.

[@chrismc410]

Jim Leonard now i feel old....MFM drives

[@apexmike849]

I *am* old ;-)

[@chrismc410]

ApexMike if you remember having to park the hard drive heads, you indeed are old. People have no idea how good they have it now.

[@jbdragon3295]

You left out SCSI HDD. I had to llf my drive for my Amiga computers. Those were even more costly back then.

[@JoannaHammond]

A modern lowl level format is really just a zero fill. It is required sometimes such as when you remove a drive from a raid array. The drive's "information" is incorrect for normal formatting and generaly will report an incorrect size of the disk. Therefore a LLF (Zero fill) will restore the drive to it's original factory settings and will then allow you to format the drive at at a high level. Not quite the same as the original LLF but still something to think about.

[@harkovdent5707]

in windows is that the "full" option in the format "dialog box"?

[@isettech]

A big difference not mentioned is how the heads were positioned. Originally the heads were positioned with a stepper, much like a floppy. With low density tracks this worked well as long as the temperature did not make big changes resulting in head to track alignment errors. With the advent of the IDE drive and more tracks, the stepper was eliminated with a voice coil Winchester drive. This had no feedback mechanism for how far the head was out onto the platter. This was a servo type drive, and like a CD or DVD player, the positioning was read directly from the disk, so thermal changes with the much smaller tracks became self aligning. A low level format would erase the servo alignment and destroy the drive. As shown in the video, a very precision set of heads with feedback on positioning was used to write the low level format. This is known in the industry as a Servo Writer. This wrote the servo track information used to number the cylinders, add sectors, and provide active servo feedback of the head position on the track. This track is designed to last the life of the drive and the integrated drive electronics prevented writing in the servo tracks. A low level format command does not attempt to write over these tracks as they can not be accurately positioned by the drive. A write in a low level format in the drive is a blind operation with no feedback of a head position drift during the write.

[@waynemacleod3416]

You can't low level format a modern hard drive anyway. (like he said) not since LBA because all track head and sector info on any LBA drive is virtual. if you try to low level a current drive it will do a high level instead and lie to you. low level has not been possible since we exceeded 504 MB per disk. -- for those who actually care

[@NetRolller3D]

In fact, true CHS addressing isn't even possible on modern drives, as the sector count changes with each physical track (more sectors can fit around the edge of the platter than near the center)

[@nneeerrrd]

Don't forget to enter C:\>park before go for making your cup of tea :)

[@Johanniscool]

Мандибрики і Цирупопики I use to park heads every time I shut down my 386, fun times.

[@JesusisJesus]

AUTOPARK.EXE

[@laharl2k]

I do the same to my car every time i stop using it.

[@1st_ProCactus]

you mean C>park

[@nneeerrrd]

ProCactus nope. prompt $p$g

[@ItsTheLondonFella]

Who remembers Norton Utilities, from back in the MS DOS days?!

[@mvl71]

London Fella I do. So much better than the built-in crap MS included with DOS. Norton Commander especially. And yes, I fucked up my Windows 95 installation with Norton Speeddisk for DOS.

[@peterlamont647]

I had FPROT...cool software.

[@iluvmyswamp7948]

London Fella i found msdos pc today by the trash and it had 512 mb ram.. 512 MB RAM!!!

[@ItsTheLondonFella]

Life is GOOD!! :-)

[@jorgequinones1723]

I didnt liked Norton Utilities, Id used to like PC Tools.

[@TheSlinq]

That clicking hard drive in the background really grinds my gears

[@ehsnils]

Another reason for low level formatting the drive in old time was to tweak the interleave in order to optimize the speed of your hard drive to your computer.

[@jwhite5008]

The sector marks are now only on one side of one plate - other plates have to be perfectly aligned with it to be readable. They are written with super-high-precision factory equipment before the plate is even put into the hard drive. Calibration parameters are stored in a small flash memory on the drive board. They are often changable with a special cable connected to the HDD, but changing them would ruin it. Bad sector list and S.M.A.R.T. statistics are stored in a special place on the disk, together with bad and spare sectors, all of which are not readable by the OS. As mentioned in the video, sector remarking makes no sense on modern drives, thus that command cannot and does not work. Low-level formatting is nowadays performed either with special hardware (such as an UART cable) or with very specific manufacturing software. It can reset things like the bad sector list - which is a bad idea most of the time, but might be required - for example if spare sectors are exhausted because of a bad cable or power supply instead of a dying disk. If you liked the explanation please like this comment so that more people can read it. Thx.

[@Lilhawke]

I've wrecked the servo boosts on HDDs before with low level formatting. That boomerang sound still haunts my dreams.

[@nessotrin]

Lilhawke What is servo boost ? As a nerd from the edge of the 21th century, I'm really curious about such a thing °.°

[@DMack6464]

Congrats on all the aqua pfps

[@alexwaite7173]

Great video. I didn't understand a word of it. But still enjoyed it.

[@helmaschine1885]

Alex Waite Me too, man. Me too. At least we got to make tea out of it.

[@KurtRichterCISSP]

Enjoy your tea!

[@JohnSmith-el5mo]

Anyone wants some biscuits with your tea.

[@JizzyLeaks]

Me during most of his Videos 😝

[@huazhu]

he saved my tea

[@TheRestartPoint]

4:34 You've clearly been saying "Megadrive" too much recenlty haha. "A Ten Megadrive"

[@zvpunry1971]

We now have more then ten teradrive. ;)

[@measl]

I think you meant "MegaBYTE", not "megadrive"...

[@Lightblue2222]

Seeegaahhh!

[@bigshrekhorner]

@@measl Listen to the video. Instead of saying "Megabyte", Nerd said "Megadrive"

[@measl]

@@bigshrekhorner *Funny! I completely missed that: I guess I'm so used to that kind of ridiculousness from people I don't even hear it anymore.*

[@pavelfara9333]

Very nice video for basic HDD understanding! Just one think to add - In VERY early IDE era with old drives using RLL method and pretty much having a controller card just slapped on to them it was possible to Low Level Format those. But you have to use a factory software and it most likely ask the firmware of the drive to do it right. For example Seagate ST-157A is such a "primitive" IDE.

[@tehf00n]

I still park my SSD drive ¬_¬

[@freeculture]

Don't let your virtual magnetic head scratch your virtual magnetic platter!

[@ianpulsford2295]

I always rewind my DVDs before taking them out too :D

[@andrewwitt5278]

LMFAO

[@sliyarohmodus5749]

@@ianpulsford2295 The microwave is an excellent tool for rewinding DVD's.

[@jarlfenrir]

Sometimes I'm even rewinding my mp3 files using pencil!

[@wildgoosespeeder]

So, low level formatting is where the user had to manually format a hard drive at the time of purchase. Then came along the IDE (precursor to SATA) where the factory does the format rather than the user. Then a controller safeguard came along to stop the low level format to be performed on the newer drives. Now when an OS install wants to format, it erases data and then marks sectors as a formatted partition, without affecting the factory created low level format. The quick option skipping the erase data process. The data still exists in the quick method, just the pointers to the data are lost. So when the OS is tasked to write data, it just changes the data missing pointers to data that needs to be written. Vsauce did a video about where do deleted files go and how overwriting data works. The analogy makes sense to an average user.

[@10p6]

This is not entirely true. Back until the mid 2000's, and still even now, you could still get custom software from the manufactures to low level format the drive in such a way that it actually rebuilds the drive bit by bit and not just replacing the sector information. Obviously when doing so it gives the user a harsh warning that "EVERYTHING' is going to be erased. The other thing is, in many situations where you have an old computer and dive you want to donate, the best way to permanently erase everything is with a manufacturers single low level format over multiple drive wipes as the LLF can provide more power to rebuild the drive than software can to overwrite data.

[@hakemon]

Byte size is possibly my favorite segment on your channel. I grab my headphones and enjoy learning something (and even if I knew something, I love hearing about it again, haha).

[@perli216]

Whats the synth music at 2:22?

[@Lethaltail]

I'm wondering that too.

[@SNESMike]

A bit late to the party, but it's Crystals by Silent Partner - available on the KZbin Audio Library: kzbin.infomusic

[@perli216]

@@SNESMike thank you

[@dev639]

I miss my 500mb drive. There was so muh space for all my floppies in it.

[@TheBrokenLife]

I have a worked 540mb drive I'll sell you. I paid $550 for it new, How does half price sound?

[@Hallkardia]

Oh my god, you are so funny

[@ItsTheLondonFella]

I remember my 20mb A590 HD, for my Amiga ... WITH 2mb of additional RAM!! I miss those days, so much.

[@peterlamont647]

I don't...I'm living the dream...got 3583 bytes free, and a tape drive to pick up the slack! If I have some heavy lifting to do, Ill cram in a 16K ram expansion. Ultra computer power! *air* *guitar* sold separately! The VIC-20 FTW!!!

[@elephystry]

Fabricio Fanfa *much

[@NoJusticeNoPeace]

I remember when I got my first hard drive, a 10 meg IDE for my XT. I was blown away that the 10 meg hard drive we'd had in our Apple ][ computer lab, which had been the size of a large microwave oven, was now small enough to fit in my hand. After copying MS-DOS, Civilization, and all my Apogee games onto it, I still had plenty of room to spare. I remember thinking there was no way I could ever fill up so much storage space.

[@jmtx.]

Thanks for an awesome blast from the past! Storage has come such a long way in my life time!

[@DanHarkless_Halloween_YTPs_etc]

3:24 - "on XT or @ machines"?? Did all Brits pronounce AT that way?

[@leopold7562]

Hell no! It always made me cringe when I heard that and in 2019, I'd have thought we'd have got past this. Clearly not.

[@robinw77]

No it's the first time I've heard it. It was always A-T whenever I said or heard it.

[@donkmeister]

@@robinw77 Thirded. I'm also British, and it was always A-T whenever I've said or heard it.

[@laurensa.1803]

You should zero your hard disk of course! By the way: I love your work Nostalgia Nerd...! Especially because I am a technically inclined person (education in computer science) and it is nice to get a laid back explanation of typical (older) aspect of computers.

[@UpcycleElectronics]

First time your channel was suggested in my home feed. First upload of yours that I have seen. Liked/Subscribed

[@frugalprepper]

DEBUG=C800:5 Those were the good days. And don't forget ESDI. A high level format wouldn't really format the drive you had to low level it or compsurf (on Netware).

[@measl]

I used ESDI drives extensively at work in the 80s (see my intersector discussion above), and I remember - _vividly_ - the pain of formatting! Formatting a 300mb drive took many hours. In fact, it got to the point where I wouldn't do it during a workday: I would start the process before going home on Friday. If someone needed a formatted ESDI/Controller pair, they had to request them in advance to allow for the prep of formatting.

[@christopherc92]

I came late to the PC scene in about 1995 so have only known IDE hard drives on the platform. I remember getting hard drives new and having to format them from the BIOS menu to get them to show up in DOS. I always assumed this was low level formatting. Glad I watched this Video, keep up the good work :)

[@TheZXKUQYB]

I remember the first installing a HD with crap load of numbers to punch in and I was like in my early teens, no sweat. It worked and I felt awesome.

[@southernflatland]

You are correct regarding more modern hard drives. Floppy disks are a different story though. Granted they are outdated, but still can be useful to nostalgia nerds and even useful for diagnostic software (including SpinRite which you mentioned) I used to salvage "BAD" floppy disks with a very strong magnet followed up by a low level format. I was actually able to salvage about 80% of bad disks this way (not salvage the data though, just salvage the disk for reuse).

[@Kepe]

Yeah this video title is horribly misleading. The video doesn't answer the question posed in the title. And how could it, if low level formatting has been impossible for almost two decades? The title is only relevant if you have an old computer laying around that has an early IDE hard drive in it. Perhaps the title should read "Why is it impossible to low level format your hard drive" instead. Or even better, change the title to something else than a click-bait headline. Perhaps "History of low level formatting".

[@deleatur]

Sorry, but this channel is called *"Nostalgia* Nerd" for a reason, bro.

[@VibbyABibby]

@@deleatur y e s .

[@douwehuysmans5959]

It did answer the question, what are you talking about...

[@DarrenCoull]

This is exactly why when drive destruction services degauss a drive, it removes these factory-added sector markers and there literally is no way for the drive to operate

[@86SVA]

What's the music?

[@gopherbone697]

The production and presentation of your videos just gets better and better all the time. I love your channel and this video was a joy to watch, like many others you've done. Thank you for the consistently awesome content!

[@dysfunctionalwombat]

Mac OS 9 (2000) Has A Low Level Format option

[@niyablake]

That's cause macs were SCSI

[@dysfunctionalwombat]

I knew that, I had a PowerBook G3, One of the last macs to have a SCSI port, although its internal drive was IDE

[@mustangrt8866]

still have that option today

[@phillipsusi1791]

Did you not pay attention to the video? That just has it zero fill the disk.

[@andycraig7734]

Niya Blake Not in that time frame.

[@38911bytefree]

With stepper motors for heads and disk expansion with heat, the limits showed at 40 probably 80MB. They need to move to another approach. Not because its cheap, but to allow higher density without worrying about disk changing in size (sligthly of course) during temperature changes. So voicecoil actuator was introduced. Head now moves to any position without any steps in between. This allowed higher density but formating based on actuator position was not possible anymore. So, they used tracks on the platters to store location information so the head can tell where it is exactly at any position by constantly reading this tracks. This also solved the thermal issues with disk expanding since heads are not referenced against a physical zero track, they just seek por tags on the tracks, if disk expanded a bit, the track just moved a bit, heads probably would pick a tag from a previous track and realize that it needs to move a bit foward or backward. Low level back in those days format needed to be done at OPERATING temperature and errors when cold where common. Another clear example was floppy drives (heads using stepper actuators) against ZIP reader using voice coil actuators just like hard drives.

[@Spolodaface]

scanlines in all videos pls

[@drunkenbusdriver]

You explained low level formatting perfectly. Any drive manufactured in the 00's doesn't need it. I still have to keep explaining Winchester drives to people that don't get it. Do you understand the sector/track/cylinder/cluster of the drive you're trying to format? Can you reproduce it and configure it in the BIOS exactly? No!? THEN DON'T MESS WITH IT!!!!!. Thank God we have a PC refresh and I don't have to deal with that now. SSDs and BIOSes that don't even have to factor that in. This "low level format" is still a thing though, I'm amazed at how many "IT Pros" still harp on it....... It is crazy.

[@NeoSeer]

The way computers work (code and software interplay etc) almost seems like a dark magical art to me . So I love it love when aspects of computers are explained concisely and with great clarity 👍🏽

[@electronicsluckydip]

Excellent video, however I'm unclear as to whether some very old IDE drives should be low level formatted or not? The BIOS in my old IDE Pentium from 1996 even has an inbuilt Low Level Format Utility... so is that utility only zeroing the drive?

[@laharl2k]

I use thermite to low level format my drives.

[@HarosOfStyx]

Laharl Krichevskoy lol I pissed myself laughing at that lol

[@wingracer1614]

"It's the only way to be sure"

[@mcrecordings]

Someone told me years ago not to format through the BIOS as this was a low level format? I can't recall if they meant IDE or SCSI drives.

[@JustWasted3HoursHere]

I like that 10MB HD for $3398. Hard to believe how cheap storage is nowadays.

[@JeffDeWitt]

And it was "The drive you've been waiting for!"

[@antoniomaglione4101]

Going further back in time, before the PC era, the 12 inch hard drives - with the grand capacity of 3 MB - like the Winchester and the Diablo, required an external hardware "disk exerciser" in order to establish tracks and sectors after an "heads change". Replacing the hard disk heads was common occurrence, say after a power outage, after forgetting to park heads before turning off, after the buildup of moisture, after not replacing the air filter, or after just being unlucky. Three megabytes would hold some monster-sized database, don't forget... Thanks for the video. You made me remember, that I purchased the MFM hard drives and the controllers as a pair, at the time...

[@WskOsc]

I bought a 160GB Seagate Barracuda in 1998, back when that drive size was brand new and I remember the guy in the shop saying "you'll never fill that, who even needs that much space?" Cut to 19 years later and it's still in my current PC and still going strong. I'm well aware of its age and don't keep anything worth keeping on it, just downloaded installers, etc. things I'll use straight away and then delete or forget basically. Wonder how many more years I'll be able to squeeze out of it.

[@TheBrokenLife]

Our first PC in 1991 came with an UPGRADED 80MB HDD and the salesman said the same thing.... It took me about a year to fill it with shareware, including the time it took to learn what a modem is, what BBSes were, etc, etc... As far as I know, that drive is still operating just fine too. Power supply on that machine died a few years ago so I can't confirm.

[@joeygreathouse3029]

No, you didn't. In 1998/1999 a 20GB drive was considered enormous. You're just wrong. Let me guess, you currently have a 160TB drive in your computer.

[@TheBrokenLife]

Thinking back on it, I think I was making payments on a 8GB drive at that time on a Best Buy card... Thing was like $400. So, yeah, no 160GB drives for a few more years after that and certainly not in a box store PC. 1.6GB is much more believable or perhaps 2004 and not 1998. Something isn't making sense. Time confuses me sometimes too with the pace that things were evolving back then.

[@n2n8sda]

Maybe he has the years wrong from memory.. Around that time I bought an 8.4gb Seagate that was fairly top of the line, died within a couple of weeks and had a pain in the arse time trying to get it RMA'd. Never used them since.

[@gabrieleriva651]

Jacob Shaw DSL and Emule filled my 80GB in 2002 in a flash.

[@cmoor8616]

What is that audio backdrop at 2:22 ? awesome!

[@BillyBoze]

You know what? I will make myself a cup of tea. :)

[@ThePhoenixArchives]

Don't forget to format it first.

[@BoDiddly]

Excellent Video! Mainly for mentioning MFM and RLL drives! I remember some of those drives weighting 10lbs. I think I might still have one in my basement somewhere... ol!

[@charlesmartin1972]

1:37 "unrecoverable" 😂😂😂 if only it were that simple...

[@PanduPoluan]

This is also why degaussing does not need to completely reset the magnetic orientation of a hard disk to make it unusable, just enough to mess up with the sector markers and the standard integrated controllers will just give up. (This is also how data retrieval companies do their magic: They replaced the controller with a fully programmable one, scan for hints of sector boundaries, and restore / overwrite the sector boundaries with their own.)

[@sbrazenor2]

I think I'm going to low-level format my SSD now... j/k LOL :)

[@pufero1]

On sdd some updates of firmaware do a low level format on the nands.

[@Ccoolty]

You should go defrag your SSD instead.

[@flatfingertuning727]

Fragmentation is an issue with SSD cards, though in a different way from hard drives. WIth SSD cards, what matters is not having files be contiguous, but having free space consolidated.

[@NolePTR]

Hence why I still run defrag software for my SSD: to consolidate free space.

[@TalesOfWar]

You're an idiot. This does nothing of benefit and instead just reduces the lifespan of the drive. TRIM negates this problem too.

[@ct6502c]

I'm still kind of confused about all this, but I'm fascinated with how hard drives work. It was cool to see this video and watch the drives spinning and see the mechanism working :)

[@SurmaSampo]

HDD Manufacturers still provide LLF tools that perform a sector remap and a zero wipe for IDE drives. I have run them as a tech and they are still the recommended first step to correct a drive that is developing bad sectors. The real change is that the tool is just a command sent to the controller that is preprogrammed to perform that action correctly and to spec. Also the "special higher energy process" used at the factory is just a repetitive write to ensure all the bit are correctly aligned as a single pass isn't that reliable. There is no depth as the bit layer geometry on a platter is essentially 2 dimensional. Plus it is possible to use atomic shadow analysis (an expensive but commercially available technique) to recover data from a zero wiped platter. I don't think you actually know what you are talking about.

[@Nostalgianerd]

I don't think you understand how to present information without getting bogged down in unnecessary detail.

[@SurmaSampo]

The problem isn't that you made mostly correct simplified statements, it is that you made demonstrably false statements that are almost completely incorrect. Saying "it is impossible to recover data from a zero fill" is incorrect so instead it should be "it is super difficult and expensive to recover data from a zero fill" which is accurate adds a whole 3 words. Your script is littered of these incorrect absolute statements presented as fact which damages credibility. Also you could use "it is super difficult and expensive to recover data from a zero fill. We will be doing another video on this amazing recover process in a future video" to springboard more content or your could use it is super difficult and expensive to recover data from a zero fill, and for those that are interested in finding our more about this amazing recover process check out the link in the description" to drive traffic to your blog or website or an affiliate to diversify your mediums.

[@askjacob]

I challenge you to find me ANY commercial recovery place that will do data recovery from a modern drive from zero fill

[@SurmaSampo]

IBM funded the research and has the lab facilities to do it in both the USA and the UK which was initially offered only at the UK lab for a minimum fee of one million pounds. It is a slow and expensive process that quite a lot of physics labs could be set up to do if they wanted to invest the time and money required seeing as the technique is over 10 years old now. If you have the money and an IBM account manager then you can get it done as a bespoke service request. Essentially the work of the read head is replaced by an electron scanning microscope that looks at the surface under the bit layer. That information is then run against a heuristic system that figures out what the bit layer configuration would have most likely looked at previously. The more overwritten the location, the less reliable the technique becomes so it is best for obtaining static data values like encryption keys. N.B Knowing which filesystem is also useful as MS file systems write system data to the inner tracks to make them faster a first boot but prone to fragmentation where things like ext write system data starting at the centre tracks then algorithmically write contiguous data blocks spaced as far apart as possible so modifications of that data can allow it to get larger without needing appends resulting in very low fragmentation and best consistent read/write speed.

[@Olivia-W]

Surma Sampo That is why any data I _really_ want to erase is: A). not written at all, no trouble there, or B). greeted by my good 'ol trusty hammer and a friendly furnace.

[@sgcs]

watching this while "low-level" formatting my hdd lol

[@TalesOfWar]

This reminds me of people who insist on defragmenting SSD's.

[@Mostlyharmless1985]

1. Do they not understand it's RAM? 2. How? 3. Why? 4. No really, How?

[@zosxavius]

So I shouldn't degrag my SSD everyday? It seems to improve performance.

[@ColdestLivewire]

ssds do show up in windows (7) disk defragmenter and can be manually defragged for some stupid reason

[@sirifail4499]

TalesOfWar Reminds me of experts who insist NTFS (some other favorite file system) never ever needs defragmenting.

[@leopold7562]

​@@zosxavius I hope you're joking! But for anyone who genuinely thinks like this then here's some advice: (a) You're not improving performance. Defragging rearranges data so that larger files are stored in contiguous blocks, meaning less searching around for the data being read - if all the blocks for a file follow one after the other, the drive can read them all in one sweep, whereas a fragmented drive means multiple head movements to locate the blocks. This isn't necessary for a purely electronic storage medium such as an SSD as there's no moving parts, hence the delays incurred are microseconds at worst. Certainly not enough that you'd notice. (b) You're actually degrading performance. SSDs have a limited number of writes, so every defrag is killing your drive. Eventually you'll be left with a useless drive, and a hell of a lot more quickly than you'd think.

[@MechWizzard]

I low-level formatted IDE & SCSI drives all the time... using manufacturer's diagnostic software. We did it to setup the P.list on the drives. The P. List (or Permanent list) is a list of bad blocks detected by the manufacturer during it's first low-level format. There is also a G.List (Growth list), this list is created by the firmware on your HDD, bad blocks are added to this list every time the drive encounters a defect when in use, its completely automatic and is usually undetectable by the OS (as that manages 'file system' errors not hardware ones). If the G.list reaches a certain threshold then you will get a S.M.A.R.T. error.

[@Puremindgames]

*goes to make cup of Tea *comes back with Coffee I AM FAIL!

[@opimentoso]

This is what you get when you physically format your brain.

[@laharl2k]

Hello FAIL, I AM ERROR, nice to meet you.

[@sparticus214]

+Thiago Pimenta I have a Linux brain so it handles high level on low level format.

[@SuperDavidEF]

Well, coffee IS technically a tea made from brewing water through ground coffee cherry pits. So, no fail this time. Carry ON!

[@BenjaminCronce]

When my IBM DeathStar HD got corrupted, caused by the glass platter expanding from heat and the head writing data in the wrong spots, I used a tool that claimed to "low level format". I don't know what it did, but it was definitely not just zeroing. In this corrupted state, I couldn't even format the drive because many of the sectors were claiming to be corrupted and unfixable. After a "lowlevel format", everything went back to working and I could read/write all sectors on the disk again.

[@pinrod1]

now you can just download a new SSD from the interwebs! Download more ram too......if you need some

[@wingracer1614]

I pirate my ram but you should always download your SSDs from a trusted pay site. Don't want those russian hackers getting you.

[@pinrod1]

Make sure you use crypto currency to pay for said SSD! Untraceable!

[@krashd]

Actually seeing as cloud storage is well established and we are now entering the era of cloud computing and streaming the old jokes about downloading hardware are coming true -in a way. One day we won't have computers, just a terminal and the computing will be done at the server side. Replace your 8K screen with a 16K one? No problem just choose to stream 16K games instead of 8K effectively upgrading yourself to a massively more powerful PC just by the click of a button... That is until CloudNet becomes sentient and instructs us to build the factories with which it will liquify us into a slurry.

[@plumbc007]

peter ,,, plzzzz for the love of god whats the synthwave tune your playing from the last half of the video??? thanks :-) brill video as usal

[@erickhauser2322]

And... all the reasons we SHOULD?

[@hoxatron6024]

So you can troll your friend.

[@moiquiregardevideo]

The MFM hard drive were using step motors to select the track. At boot time, the step motor is commanded to go to zero. Since there was no limit switch, the just let the head bang on the mechanical limit enough times to cover the worst case ; that is, if the track at boot time happened to be the highest one. The IDE drive replace the step motor by a "voice coil". The head is free to wander anywhere in between tracks. The hard drive can contain from 1 to 16 heads. Head: either one or two per disk Disk:1 minimum, may have 2 to 8 disk all rotating together Track: a full revolution of the disk Sector: 512 bytes data, each track contain from 8 to 128 sectors The total size of a hard disk is the sum in bytes of each sectors, so it is always a multiple of 512. A track can only contain an integer number of sectors. It is possible to write more sectors for the tracks located close to the edge and less sectors near the center. There is always waste for the space that remain of each track. More precisely, we know for sure that 1 to 511 bytes is wasted on each track, this space can be used to provide extra redundancy about the track number. Note: The head is 99% in read mode, When we send the command to read or write a sector, the controller apply DC current to the coil and attempt to detect on the fly which track/sector the head is located. Seeking a track/sector Once the head is approximately at the right location, a fine tuning loop can sense and correct the exact center of the track. A very fast oscillating current in the voice coil attempt to maintain the head dead center. Meanwhile, the controller is reading every bit on the track and wait for the rotation of the disk to reach the location where the sector is located. Write operation: Once the beginning of the correct sector is detected, the head immediately switch to write mode and start writing each bit (512 bytes * 8) and other information (checksum, stop signal and x bits of repetitive "1010..." for clock recovery. The head immediately switch to read mode Read operation The read operation is simpler since the head is always trying to read. The difference is that the data is captured inside the internal buffer of the IDE drive (which was 8 Meg RAM for more than 15 years, the only model of static RAM that remain available on the market after DRAM won the race for size and speed). Low level formatting difficulty Low level formatting is possible but very risky because the head has no reliable way to know where it is located on the disk. If a single track need to be fixed, then it is possible to detect the previous track then move to next track then backup and try to write the broken track hoping that the head didn't move too much. Small Random variation in sector position The write operation described above has some timing variation. The sector will not start exactly at the same falling edge of the start bit. The variation can even move up/down by many bits. For that reason, there is a dead zone between each sectors. It is expected that each time a sector is rewritten, it will wiggle a little in position inside the allowed dead zone. Other problem is the irregular rotation speed of the disk which make each sector stretch or compress. The dead zone must be large enough to allow the worst case where each sectors on the track happen to be stretched to the maximum limit allowed. The major cause of irregular rotation speed is the low resolution of the step motor that drive the rotation. It is typical made of 3 phases that repeat 3 times (9 coils total). The electronic apply current to each coil is sequence producing 9 locations where the disk accelerate per track. The PLL that generate the clock/data rate can also wander up/down a little which has the same effect as changing the rotation speed. It contribute to compress or extend the physical length of the sector. Failing FAN, the biggest enemy The worst problem for IDE drive is vibration. For example, when the fan on the CPU start to make that loud noise because the bearing is loose and the fan blade move out of the rotation axis a little. Such annoying sound also produce strong mechanical vibration that we can feel just touching the computer case. The hard drive perceive these vibration as treat with the risk that the head overwrite the next track or more. The hard disk controller may wait many disk rotation until it sense the centering loop stable enough. Once a sector is started to write, there is no way to stop it. The electronic will continue writing all the 512 bytes. It is only after it complete to write the entire sector that the hard drive can try to sense the position of the head. If it finds that the head moved to a different tracks... too bad, it is possible that the sector was not written completely and that the end of another sector on a different track got corrupted. Laptop drop, even worst The hard drive in laptop have an even tougher job to ensure writing a sector on the right track. The problem is that even if the laptop is gently moved to the table, the exact time when a contact with a hard surface occur, it can easily be equivalent to an acceleration of many G (many times the equivalent of the force of gravity). A sudden shock which can be seen as a square wave on a scope, contain a large number of odd numbered harmonics. Dropping the laptop by 3 inches is equivalent to attaching firmly the laptop on a vibrating table with frequencies in the order of hundreds of Hz. Some laptop hard drive incorporate a motion sensor (accellerometer, like we have in all smart cell phone) which interrupt writing a sector at anytime and restart writing again the same sector when the head recalibrate to the correct track.

[@kewlztertc5386]

This was funny, 8 minutes of absolutely useless information. The next video should be "why you should never put a floppy disk, into a CD drive" then explain how we don't have CD drives anymore so nevermind. Lol

[@Nostalgianerd]

But yet.... you watched it. Thanks!

[@gageboyd5504]

Thank you so much for introducing me to the fever dream that was that MS DOS 5 upgrade advert

[@JonnyRocksAW]

For real that was a gem.

[@kevinfishburne]

For modern users: "sudo dd if=/dev/urandom of=/dev/sdx" where "x" is the drive letter, starting with "a". "sudo blkid" will list drives and partitions. I don't know what Linux would do to an old-format IDE drive when issuing such commands. Maybe brick it, maybe secure it.

[@thepumpkingking8339]

Question: I understand a formatted H/Drive has an X defined amount of sectors on each disc platter, depending on the size of the disc. An SSD does not have platters. So is it structured / formatted as if it does have physical, platters or is structured more like a stick of memory, for read / write speed.

[@AhmedKMoustafa2]

What is the difference between filling sectors with zeros and filling sectors with zeros & ones ?? why the second one is slower ?

[@BryonLape]

When HardCards started to appear for the Tandy 1000 units we were using, I used DEBUG to get into the firmware on the controller card to low level format. SpinRite was a god-send in those days.

[@robehickmann]

While you don't want to low level format the drive, there is still value in scanning the whole thing. This can allow the drive to identify bad sectors and replace them. That's what SpinRite 6 does, with some custom error recovery - assuming that you have a computer old enough to run it.

[@tredfxman]

And how do you expect me to get rid of clever boot sector viruses, like example the ones hiding in Windows.Old from old Windows installations, Without a Low Level Format... ?? 👍

[@jonasferraz]

Delete this folder using Shift + Del

[@alexramos2568]

What about the low level formatting that was available in some old mother boards, what was it ?

[@andreassjoberg3145]

Last time I LL-formatted a disk, was after the IBM Deskstar-crashes. They had a faulty controller software, and filled up the error proms and crashed. The fix was upgrading the PROM of the controller and then doing a low level format with the diagnostics software, and then doing a proper format.

[@sailbatten2056]

I low-level formatted my Fujitsu 3TB and the format took over 20 hours. And it performed perfectly, and during the format found and remapped bad sectors. (?) Did it somehow decide to just spin there writing zeros only and report nonsense?

[@The_Vanished]

I always wondered if there are heads and cylinders in hard drives, do they need a block, connecting rods, and pistons?

[@DragonProtector]

man those commercials are so awesome compared to today

[@rickperez8044]

I don't understand. I thought the debug routine at C800 was a bios extension residing on the controller. Without a controller, there would be no bios extension to begin with. This should not even be a possibility. What gives?

[@WizardNumberNext]

you still can low-level format a lot of Hard Disk Drives, but this does not apply to consumer Drives all enterprise grade drives can be low-level formated (SCSI, SAS, Fibre Channel) this is to accommodate different block sizes like 520B and 528B for instance

[@atfsgeoff]

I've only ever had to do a "low-level" format on one drive, it was a SCSI drive from a mid 1990s unix server that had been in continuous operation for 20+ years. It lost its cylinder-head-sector boundaries so they had to be redefined, and then an earlier binary image of the disk was loaded back onto it.

[@Twinrehz]

Would it be possible to write software that would actually allow you to mess with the hard drive controller and get it to low-level format, and would this make data recovery impossible? Or is professional data recovery methods today so sophisticated that this would hardly be a challenge to them?

[@solidwire]

That's an awesome idea. Really feel like a cup of tea now that you mention it.

[@jkp910]

I just low-level formatted a bunch of HPE SAS drives last week. Most Enterprise-class drives have a 520-bit logical sector layout, which is incompatible with most RAID controllers. You have to low level format the logical sectors to 512 to be properly utilized.

[@jackdeath]

I don't know. I picked up a dead WD drive from a yard sell, low level formatted it, and it works fine, now. And some drives were notorious for holding on to virus until I low level format them. And while, yes, it's an extreme measure to put the drive through, it's sometimes your only option.

[@Palooka37]

Just popping in to say how much I'm appreciating the old content that the algorithm keeps throwing at me. :)

[@connorshaw312]

User: *tries to low level format drive* Modern drive: “what the fuck are you trying to pull on me? Fuck this I’m zeroing all ur data.”

[@niyablake]

You still can low level format SCSI drives. not SAS drives though

[@r4microds]

Holy crap! Spinrite!!! I remember using that software on a CD mule times to save old drives with bad sectors. As a kid that had no money it saved me my laptops dying drive by actually scanning. Flipping and re scanning and then doing something to block out the bad sectors while also recovering damaged data. To this day I'll still remember it as that program that saved my laptop.

[@oturgator]

Do you know why they made the Mainboard manufacturers to remove the Low Level Format option from the BIOS?

[@francoisleveille409]

On my first PCs, I used only SCSI drives because they performed so much better and were much more reliable. Low level format was a built-in functionality on all Adaptec SCSI controllers I have owned and was accessible through the BIOS of the controller (CTRL-A). It did save a few hard drives that were otherwise considered dead. Nowadays, low level formatting on SATA drives can effectively render them useless. Obviously, it's completely pointless on SSDs as it's not even a magnetic form of storage.

[@MaxContagion]

if you mixed up a hundred controllers and a hundred drives what would be the easiest way to sort them out? or basically if a drive is already formatted for one controller can it be low level formatted for another controller?

[@hqt00]

I did this once on an old IDE drive that was giving me issues. I apparently got lucky (or the drive protected me) with the data table on the drive as it came out working after it was done when before that it was not. That, however, was the only time I have felt the need to LLF. Everything else just 0 it out and this will solve most common issues assuming the drive is not too damaged.

[@JohnSTF72]

Isn't writing Zeros (empty data) as effective as low level format for erasing your data?

[@measl]

+JohnSTF72 It's close enough. If NSA really wants to know what was there they can still recover a great deal of it by scanning the edges of the tracks as a stream, and then working out what was in there. The process is slow, tedius, and unreliable, but I have seen them do this with one disk, again, in the mid 1980s. Today's IDE drives may have their tracks too close to each other to do this anymore, I'm not certain. The _ONLY_ way to be certain you have permanently erased a drive is to literally destroy the internal substrates by drilling the drive and then acid bathing it (routinely done at large companies).