This reminds me of the day that I met my first "PC" hard drive. It was in 78 or 79 while at school. We heard that a company was going to be in the city to do demonstrations of a new type of hard drive for personal computers. A group of us nerd girls got in the car, ditched classes and drove 100 miles to the city. We went to a seedy motel room. We were more excited to see a new drive technology than were afraid of something sketchy (times have changed). There was a guy in a room with a briefcase filled with electronics of some sort and a full height 5.25 drive connect via a ribbon cable. He gave the talk and ran some unspecified program/demonstration to show off a 5MB hard drive that none of us could hope to afford.
Oh this reminds me of the day I met my first guy in a seedy hotel room. He gave the talk and ran some demonstration to show off nerd girls none of us could hope to afford.
I learned A LOT! I didn't realize a half hour passed and when Adrian mentioned a part 2 I was thinking, "Part 2, why? It hasn't been that long!" I was so wrong.
I have the feeling that part 2 will be major disappointment, its a miracle if one of those drives still work but you should not be surprised if none of them does anymore. These things could fail easily, even back in the day they are too fragile and /true/ repair requires special dustless station and parts that don't exist anymore unless you have spare drives to cannibalize from...
@@freeculture I bet you at least the ST-225 will still work. Those drives are damn near indestructible. I have a pair from 1985 that I was able to low-level format last year with zero defects.
@@freeculture I bet you that the Micropolis has a sticky bumper or a stuck brake but will work, I bet that the ST-225 works no problem, the ST-4051 works, and the PTI drive is dead as a doornail.
Great video! 😊- I worked in a computer shop in he early 90's and remember fitting these. I remember that one of the main causes of failure was people not being bothered to 'park the heads' before they turned off the PC, thus the drive stopped with the heads resting on the drive platter. Any movement of the PC would likely result in disc damage and data corruption. New low level format, new bad sectors mapped and back to the customer.
I find it so fascinating to learn about the history and evolution of these legacy devices. It makes me appreciate the engineering, both mechanical and electronic, that much more. Thanks Adrian.
God. I remember having to deal with the old HDD interface cards. You had to go into debug and it was something like C800G (that is not the correct address, but it was something like that) to execute the code on the HDD interface card to do things like low level format, mark bad sectors, etc. IDE was a lifesaver and a hair saver! ^-^
@@rogerjones8809 Yeah! And the address changed depending on who made the card. All text too, so no EZ mode. You had to actually have a clue what you were doing. Glad to have moved past that. It was no fun at all.
I didn't see this but added the correct info as others had replied to in your comment above.I used to install the ST-225's into IBM XT's for Chemist shop systems in South Australia back then. Ahhh Memories
This is awesome! Really enjoyed the video and can’t wait for part 2! I’ve always been extremely fascinated by MFM hard drives and have been collecting and repairing them for quite some time now. I always use period correct MFM drives in my vintage PCs for the authentic sound, speed and feel. I would love to see a video on RLL controllers as well!
Oh man, I miss when drives were still heavy enough to be weaponized when dropped by accident. Still miss not needing an LED to tell when files were being accessed as well. Love seeing the progression and advances in engineering.
I don't miss them because they were very fragile, heavy and fragile is not a good combination... I'm still using mechanical hard drives right now and i still treat them like handling a bomb when i have to move them for some reason, from habits of those days. I wouldn't dare do things like what Adrian did in this video and back then the things were darn expensive too! I remember some of the manuals would show/tell you could not run the drive upside down or vertically or at some odd angle for example. Yes, there was a time hard drives included printed manuals, i still have mine for the ST3144 family which includes de ST3096 i own (and i actually got the ST3144 later from somebody who didn't want it as it had some bad sectors), one is 80mb the other 120mb iirc, both have their own LED and are 3 1/2", never saw drives with one after these.
At least in those days when someone stole your computer, you just had to wait about five minutes, then walk out the door, and the first person you met outside who was out of breath, was the thief!
I'm glad you mentioned that. At a sales/repair shop in 1990ish, my tech partner is working on someones XT/8088 writing machine. (Lots of authors in this area) At one point in time (1985?) author buddy upgraded to a 20MB full-height 'Winchester' as they say. As buddy is pulling the case top forward, angle up and off, the chassis itself angled down sharply enough to drop the behemouth drive right onto the top of his toes! It wasn't screwed in OR connected to power and data cables. Well, this was a gruesome sight! Tech buddy squeeled like a dying baby rabbit. To make matters much more unpleasant, our tech buddy is one of those who wear open toe sandals year-round (In Canada no less). Let's just say OSHA didn't approve said sandals... It completely shattered two of buddies toes, pierced the skin and issued a healthy flow of the red life-juice. We sorted him out, and got him an ambulance. It is true as a Motherfuck: He never walked right again.
G=C800:5 is burned into my memory from many many (many!) years ago. Still have the ST-225 I purchased in 1987 (along with the WD controller), still works today! Amazing engineering.
You forgot one type of controller card, ERLL! Back in the day, my brother and I ran an AMIGA BBS out of Tacoma, WA called "The Missing Link BBS". We originally started the system using an Amiga 500 and an 80MB SCSI drive. The Seagate 80MB SCSI drive was so expensive that it literally cost my brother 2 months pay for the hard drive! In order to expand our BBS file capacity, we converted over to my IBM XT Clone, using MFM drives that were much cheaper. We had 2 120MB MFM hard drives, but that wasn't enough. We did some research and came up with a 16 bit controller card that used ERLL! ERLL doubles your capacity, 25% more capacity than an RLL controller. With this we were able to keep the board going until 1994 when the PC graphics and software availability overshadowed the Amigas. ERLL did work with MFM drives, which is what we used and they were extremely reliable. We never once had a hard drive failure during that entire 10 year time the BBS was running.
@@kynkokytsumi1931 Good habit. Not strictly needed if you were not actually moving the computer or drive, but you never know. Thankfully the IDE drives would park on their own when losing power.
@@freeculture the main problem was head crashes when power was lost or switched off as the stepper motor can't place the heads on the "landing zone" unlike voice coil drives which as you say automatically goes to the landing zone when power is lost. On stepper motor drives the head lose the ground effect and the head then scrapes the surface of the disk. Happened to a friend when he had a power cut and lost data due to head crash.
Watching this reminds me of working in a clone shop in the early 90's. When we would get a new batch of hard drives we hadn't seen before, we would have to call the manufacturer's 800 number and request a fax with the latest head/cyl/track values before we could use them. I still remember seeing the first motherboard with an integrated IDE controller that could "auto-detect" the values. I still remember the Conner CFS420's 826 16 63 values all these years later. I must have typed that in hundreds of times while I worked there. Great video as always Adrian!
I used to do a lot of Netware, from 2.0a on. The earlier versions of the MFM controller driver only supported the geometries in the BIOS ROM. Soft configuration schemes were not supported at all. I became quite proficient at hex editing ROMs to add weird and wonderful geometries.
I started out as a student PC tech at my university in 1989. I was so glad when, in the mid '90s, IDE drives had become so ubiquitous that I could stop lugging around my "hard drive bible" ... a fat book with all of the heads/cyl/track listings for hundreds of different models of disk. IDE was a massive leap forward.
I had to re-watch your "This thing can backup and emulate MFM/ST-506 hard drives for your retro computers" video from August 2021. It is indeed, in my opinion, the most powerful tool you can have to analyze and fully evaluate (and even image) every one of those hard drives. So cool that you present so much on this subject! Keep this up, it's truly my favorite Adrian's Digital Basement topic!
Great video! I worked in a local computer store in 1989-1990 and this brings back great memories. Some things that might help 1. The first number in these seagate drive represents the form factor. 4 - 5.25 full height 2 - 5.25 half height 1 - 3.5 2. Full height / half height refers to the bay size. So the lager 3.5 drive would still be considered half height. 3. The following numbers represents the un-formatted capacity. St225 - 25 meg un-formatted /21 formatted St251 51 meg un-formatted / 42 formatted 3. To setup 8 bit controllers Run: debug G=C800:5 I still remember this 34 years later.
When I was a teenager working for a large distributor of Seagate, WD and Miniscribe, we did extensive testing for weeks continuously in a CRC bad sector tester. The machine we were using was the very same one that prints the bad sectors on top of the drive. We failed to find any MFM drives that was made less reliable by running on an RLL controller. After these tests, I never ran a drive with MFM scheme again for my own use at least. The increase in capacity was almost universally experienced as 1.5x
I think it just had to do with warrenty to cover the manufacturer's backs, I never understood how a RLL controller would stress a MFM HDD more than it should. The way I understand how it works is that RLL encodes data more efficiently than MFM does, nothing less nothing more.
I can add that back in that time, it was also common to add such drives to Amiga (with famous ALF toolkit). And everyone I knew (including me) bought a MFM drive and used it with RLL
@@Alexis_du_60 We did prove that in the sample sizes we had. Over the course of one summer we probably put 50 drives of different MFM generations under those conditions. A couple failed, but that was well within the failure rate at the time, regardless of encoding method.
@@videowatcherdeluxe It is as reliable as drives were at the time. In the early 8 or 16-bit days, I never lost a drive. Now, of course, even though drive technology is basically magic, I have lost a few, but only over the decades since.
I can remember the moment I reconfigured my early 90s hard drive controller from MFM to RLL. My 40MB drive was transformed into a 63MB drive, just like magic! I can't recall the brand of drive, but it seemed to work perfectly fine with RLL encoding.
The standard IBM PC interface used IRQ5 for these original MFM controllers; but when Tandy produced their PC Clone Tandy 1000, they deliberately "proprietized" the interface and changed the interrupt. Standard, off-the-shelf controllers wouldn't work despite bus-level compatibility; until I saw a magazine article with the fix; take a screwdriver and scratch out a trace on the controller card. That changed the default interrupt the board used to one the Tandy could talk to. And, yes, it worked 😁. It was only when I took my 1000 to a Tandy repair center for a warranty video chip change out that a store manager wailed at me for having a "non-Radio Shack hard drive," and I told him, "yeah, and it has nothing to do with the character generator ROM issue." He hushed, changed out the ROM, and I left happy 😁
GAWD, talk about bringing back memories. I've installed countless ST-225 drives into IBM PC's and XT's back in the day. Even remember using DEBUG to jump into the ROM address on the controller card which the command was G=C800:5 or telling Debug to GO to address C800 Byte 5 to start the routine in ROM on the card, WHICH then brings up the menu and you go from there. Keep up the good work there Adrian. love watching your vids. as I LIVED in the era of Tandy TRS-80's (I had one) and yeah became a digital engineer.
I think the reason that RLL was never really used on floppy drives is because, as you mentioned, RLL can be difficult to decode properly if you don't have an extremely consistent disk rotation speed, and floppy disks can vary a _lot_ more in rotation speed compared to hard drives. If the exact conditions of the disk itself change at all over time, an RLL signal written to the disk could become basically unreadable, and with floppy disks that could easily happen due to a little bit of dirt, or a slight stretching of the (plastic) magnetic media, or just temperature changes, etc.
The early on the MFM ST225 was a 3600 RPM stability failed version of the RLL ST238 . The early RLL controllers could not RPM synchronize well with drives that did not meet stringent requirements . A completely separate company modified on of these modulation and controlling mechanisms to make things called video image files where each track of the multi platter had stored on it half of single NTSC monochrome video frames using a pulse width modulation technique . The ST225 quality version would create NTSC image flagging whereas the ST238 quality version had way less flagging . A likely reason why ST225 were more prevalent is because the ST225 was easier to get information out of as the spindle would ware out , which without which a later generation controllers could compensate for would send a ST238 to the scrap heap .
@@donwald3436 I dunno about that.. I can't find any real details on the media encoding used on those sorts of devices, but just using RLL would not account for the large capacity improvements by itself, and they also used flexible (plastic) media in non-sealed environment which would have similar stability problems, IMHO, so I wouldn't be at all surprised if they still used MFM or something like it too, for the same reasons. The main reason for the higher capacities in those drives, and successors like the Zip drive, etc is, I'm pretty sure, a combination of higher coercivity media and high write currents allowing for higher magnetic field density, along with more precise head positioning allowing for many many more tracks on the disk (poor track density due to imprecise head positioning was really the largest source of "wasted capacity" in most traditional floppy designs). Though it is possible that the LS-120 "SuperDisk", which had an optically-coded reference track may have had precise enough real-time timing information to make RLL also be feasible for that. I can't find anything that actually says one way or another, though...
@@foogod4237 I remember (hearsay!) that some folks on the FIDO hardware groups mentioned that not all MFM HDDs were able to be used with RLL. Specifically the ST-4096 seems to have had overheating issues with the driver electronics.
@@securityteam Electronics problems wouldn't really make any sense, IMHO. There's no difference to the drive in the sorts of raw signals it's being asked to work with. The only difference between MFM and RLL is how the raw on-off patterns are translated to data bits by the controller. The signal timing and frequencies are all essentially the same from the drive's perspective. There were some MFM drives which couldn't reliably be used for RLL because their spin rates were not reliable enough, so the controller sometimes had a hard time reading back what it had written because the bit timings varied too much from one part to another (precise timing is much more critical to correctly interpret RLL codes).. As AI T mentioned, that was the main difference between the ST-225 (MFM) drives and the ST-238 (RLL) drives.. They were basically exactly the same drive, it's just that one was tested and certified to be guaranteed to have a rotation speed which was stable enough for RLL, while the other wasn't guaranteed (but many people did try to do RLL with ST-225s anyway, and it did sometimes work).
Brilliant Adrian - I have fond memories of these type of drives, particularly the ST225 - my very first PC clone had one of those! Very informative as usual and I've learnt some more about these things!
Had fond memories of this when I was young and had my 286 with a 35MB MFM HDD, nice to see and learn finally how they worked, currently my oldest system is a 386 DX40 hope I come across a working 286 or XT still with One of these. Thanks for the new knowledge, i must say you're the best in explaining the concepts and theories behing these old tech. Thnx again.
And a video about ESDI. Had a XT clone with a full height 5¼" 650MB Micropolis and a half height 350MB Maxtor. Almost 1GB of disk space and MS DOS 3.3...
Adrian - this was a ton of fun - I haven't heard these acronyms in years, thanks for the drive interface rewind. I had forgotten that ATA was basically ISA on a ribbon cable. So many alphabet soup acronyms I used to see every day booting my computer... ATAPI, IDE... so many acronyms browsing PC advertisements in the 90's... Ultra DMA, Ultra ATA...
My ST-506 hard disk was connected to an Z80 computer using a XEBEC 1410 controller around 1982. I was working on a Z8000 based system at the time using the ST-506 and XEBEC 1410 controller so just ported one over for my 256k CPM 3.0 system. Was equivalent to 5 double density / double sided 80 track floppy disks - WOW! I later used the Adaptec RLL ISA bus controller to update my MFM drives for more storage. Never had a problem with the RLL encoding.
I remember the first MFM hard drive I installed on my IBM. It certainly was a learning experience. ( no internet yet and BBS's would take a few more years to mature ) From memory, the hard drives bought after the 80's also had the head and sector info along with the bad sectors printed on the top.
Sorry for laughing at the "finger acrobatics" 😁. No one saw me. Thanks Adrian for yet another very interesting video. I still remember my 3.5" HDD, possibly about 200MB, "mounted" in a plastic ice cream box outside my Amiga 1200. I had to drive the computer and harddrive with a rebuilt AT-PSU. The 1200 was nice by giving us an internal IDE-connector, compared to the 500 where one had to use the left side expansion port for such pleasures.
Brilliant, Mr Black. I do love thèse walks on memory lane. Especially because I wasn’t a pc lover at that time so I’m on full discovery mode. But it helps a lot to understand why we are where we are now. So, please, keep them coming, RLE included 🙏
I enjoyed your post on MFM hard drives. I am an older adoptee of computers, not until 1988 at the age of 37 did I acquire my first a Commodore 64. In the early '90s I stated rebuilding and selling used computers having self taught myself to do so. Most st first had 10 or 20 meg MFM hard drives with most being full height drives at first. Later on the hard drives were 20 or 40 megs in various configurations up to 80 meg total. I usually left the first drive uncompressed and used drvspace or dblspace on the second drive to get more storage space. I used a lot of different brands and capacity MFM drives the largest being the old full height Maxtor and Miniscribe 80 mb drives, the top covers being cast with the company name in large raised letters on top..A work of art and not cheap either. One thing I did was I tried to buy the hard drives with the least number of errors. Also hard drive setup was time consuming sometimes taking hours on a very large drive. Another thing I did for some of my own drives was to format n MFM drive with a RRL controller and then drvspace the drive. On a 40 meg drive I would end up with somewhere around 90 megs of space with the conversion losses causing you to lose theoretical space. It made for a slow drive with lots of clicking and drive activity light going nuts but it worked.
8:36 I suspect you mean "binary encoded" not "binary coded decimal" for the head select signals. BCD would require 5 bits to get to 16 heads. :) Really cool video, and very excited for the next installment to see which ones will work. I know I have my predictions!
TIL Shugart became Seagate. That's one of the things I love about episodes like this one! At a previous job, one of the other techs told me of some trouble he had when he was working on a notoriously sensitive test fixture belonging to one of our customers (if it ever moved on the floor even a quarter inch, a factory tech had to come out and recalibrate the fixture). It was driven by an ancient IBM PC with full-height hard drive, and despite numerous requests, the company saw no need to upgrade the PC with a newer model. That changed one day when the tech was downloading test results onto a 5 1/4" floppy (this was in the late '90s). He suddenly heard a CLANG so loud that he thought something had fallen off the PC shelf onto the floor! The company (reluctantly) sent out a replacement PC, and you could tell that although it finally had a 3 1/2" drive, it was literally scraped together from whatever spare parts they had laying around!
I actually have on old ST-4051 on my shelf that I've had issues pulling data off of! Very interested to see your approach. And you should definitely make a video on RLL! A lot of pre-IDE disk technology doesn't get a lot of attention and it could help people that run across these older drives so they don't accidentally get written off as dead.
One of the first PCs I had is one that I built from scraps at an electronics recycling shop in the early 1990s. Of course nothing was properly labeled, and there were a lot of junk parts - but I learned a bunch about hardware. I ended up buying an RLL controller assuming it was just a regular MFM (same connectors means it's the same thing, after all -- right? Ha!) and was super surprised to find my 10MB ST-412 formatted out to 15MB! I used RLL controllers on all kinds of "MFM" drives from that point on, and never had an issue. It wasn't until much later I learned that using an MFM drive with an RLL controller "wouldn't work right." Ah, those were the days... :D Thanks for keeping all this history alive.
I installed an ST225 with an Amstrad PC-1640 which I reconfigured to work with a Sony TV. The hard disk was a little noisy but worked very well with an MFM interface card. I installed it outside the case so I could retain the twin floppy drives, for some strange reason that I can't remember.
I didn't start to work on PC's until I was a young adult in the mid '90s. So while I saw these old MFM hard drives back in the day I never really learned about them. So I find your video contact about this subject fascinating and educational.
Thanks for this video. I remember these drives from the beginning of my career. By the time I planned the purchase of my first PC (in 1989), I was trying to make sense of the specifications for MFM, RLL and ESDI hard drives. The decision ended up being made for me by the system builder, who put in a 90MB SCSI hard drive that worked fine once the computer finished booting, but that old 8-bit SCSI controller took 3+ minutes start up. The SCSI card and hard drive were replaced in 1991 or '92 with an IDE interface and hard drive. (That computer was literally a bargain basement build, and in retrospect, I think it was pieced together as a conglomeration of questionable purchases previously made by the builder. It was built on an ELT-200B 80386DX motherboard that used hard-to-find, expensive SIPP memory modules and proved to have a lot of strange compatibility issues with all sorts of components and peripherals.) I would love to see a video about RLL hard drives, and also ESDI. It seems to me that these were two technologies that were only around for a short time and didn't make a lot of splash in the market, although I know ESDI was better known because IBM used it in at least some of their PS/2s.
This video hits FAST and HARD with the avalanche of knowledge of so many things I've wondered about and were all answered. I feel like it's in your very top tier quality videos. Kudos. Also, I had a ST-225 drive inside an external Macintosh Hyperdrive FX/20 enclosure and it failed while I was changing diapers on my then baby son midpoint into copying critical data on diskettes. Thankfully I found the files in a Classic 1 later.
Coming from an Amiga 500, I didn't own a computer with a hard drive until well into the 3.5" era, ( 386 PC). So I skipped over all the earlier shenanigans. This video was really informative. 👍
I used three MFM-drives on my Amiga 2000. Two of them were outside the computer and I swaped between them. The one in the Amiga had Workbench. Used one drive for games and the third to utilities.
Great video! My first PC was an AT&T PC 6300 with a 20mb RLL encoded hard drive. The only thing I have left of it is the 5.25" floppy drive. I miss it. Looking forward to part 2 and some Coleco Adam content.
I too missed having an AT&T PC 6300..... so I hunted a nice clean one down and it's safe and sound on a shelf in the basement. I even managed to track down the matching color monitor so now I have the option of using monochrome or color on rainy days when I fire it up for some nostalgia. Mine came with a trusty ST-225 factory installed hard drive in it, some day I'd like to track down an original Olivetti hard drive for it too.
Besides being a trip down memory lane, this was an eye opener. Many of the early stuff I was learning had no one to lay it out like you just did. AWESOME JOB!! thanks .... signed, 1980's engineering student.
The memories of putting my first 5mb full height hard drive in my home soldered XT. "Debug G=C800:0005" is the first thing that comes to mind. Being able to use PCTools and Turbo Pascal without switching floppy's was a whole new experience.
My first harddrive was an indeed a 20MB Seagate ST-225 MFM drive - but I used it for my Amiga 500. This was actually a low-cost solution at the time, using a PC-MFM controller card made by OMTI and a simple adaptor card (two sockets for the Amiga and the OMTI and a few logic chips). This adaptor was built from a magazine article at the time and it worked like a charm. I remember its performance being about 300 KB/s which was enormous at the time. Many programs weren't larger than that so loading times were quite snappy. The only real downside being that one couldn't throw games willy nilly on it as they either used copy-protected floppy disks or were simply not behaving very well (they used all RAM, thrashed the OS or used their own file system, IIRC). At least most of the games I had. For example, I remember being super annoyed that I still had to swap floppies when playing Ultima V. When I saw the screenshots of Ultima VI on PC in its 256 color VGA glory I sold the Amiga immediately for a good price (I had put the 500 mainboard into a PC case, the keyboard in a separate shell with a spiral cable and this contraption used a regular PC power supply with that big clunky switch). That was in 1991 and since them I'm happily in the PC camp.
I love hearing from folks whose first computer was not an IBM. Actually I started on IBM mainframes before PC's even came out. My first desktop computer for work was TRS-80 Mod 4. Yep, we put hard drives in them.
@@crankshaft3612 Oh the Amiga wasn't my first computer but rather the fourth... (ZX81 -> Atari 400 -> Atari 130XE -> Amiga 500 -> Lots of PCs made from components...)
Loved the introduction. I've seen and used all these types over the years but didn't put together their genetic lineage and evolutionary family tree before. Fascinating!
Back in the late 1980's I was the guy if you needed your seagate drive repaired, or at least have the data recovered. I worked on other drives as well. We had no clean room, and really when these drives were bad, a cleanroom was the least of your problems. I had one st225 that had a head crash that was allowed to run for a coup[le weeks, so the first step was to blow out the ounce of aluminum dust from the drive, then to carefully wash it completely clean before attempting the recovery. In that case I was able to recover the one file the user was asking for, after I showed that I was not going to be able to pull a full recovery because one platter was scrubbed free of media. With another drive recovery for AMD, they had a chip design on the drive, and no backup. AMD got lucky, I was able to perform a repair of the drive's electronics, and pull their data over onto a better drive, and recover their chip design. It was a rll drive, which were the worst of the drives because rll drives tended to loose data long before the drive "went bad". RLL drives used the SAME controller as the MFM, so I don't know where you are getting that, you are completely wrong on that... However there were many different oddball controllers, so perhaps you are talking about one of those... But the standard western digital controllers for both the xt and at would transparently operate both mfm and rll drives. AMD got lucky, I was able to perform a repair of the drive's electronics, and pull their data over onto a better drive, and recover their chip design. I stopped doing drive repair when the IDE drives came out as the "smart" drives wouldn't even spin up if they thought they were dead, so there was no mode I could run them in to do the sector reads I needed to do, and the servo media readers were not yet available. Interestingly, I had a 100% success rate with the MFM drives, 60% success rate for rll... that went to 30% when IDE came out. It was an easy decision not to work on IDE's. I had some success, but it wasn't enough to justify the labor, as I couldn't charge for a failed repair, and raising my rates 300% to make up the difference didn't sit well.
" RLL drives used the SAME controller as the MFM" Yes, thanks for confirming my memories of that. Perhaps the drives I used (can count on one hand) were those Western Digital ones, because I remember MFM and MFM/RLL being exactly the same hardware, I was under the impression that "RLL" was really just a format.
Both so-called MFM and so-called RLL disks (or rather controllers) are using MFM modulation and RLL encoding. The so-called MFM disks are using RLL(1,3) encoding, while the so-called RLL ones are using RLL(2,7) encoding.
Can't wait for part 2!! I'm weird and used to collect old HDD as a kid. Had quite the collection of these beasts! Luckily I still have a stack. Something about hearing them spin up, low level formatting... it's relaxing to me. I have recently found a pair of Shugart 1004 8" HDD that I would LOVE to get going someday! Just no idea how to 😅
My first HDD was a 20 MB MFM drive that was dropped while left unparked and was discarded as completely unusable/dead. I was working in the IT dept at a government office as a student (summer job). I was able to get the drive for free (one man's trash is another man's treasure). I got myself a used MFM controller card (still expensive at the time), and spent countless hours running tools on the drive to mark blocks as bad. I think about 10-15% of the drive was unusable, maybe more, but it survived and I used it in my 1st PC for years as I went through school. Then my sister used the PC for a few more years until the motherboard died. The HDD was still cruising. This thing was a tank.
I had a bit of experience swapping controller cards back in 1988. At that time I bought a Miniscribe 3650 which is a 40Meg drive and 8 bit MFM controller using it for several months. Later a friend of mine bought a Miniscribe 3675, a 60Meg drive with an RLL controller. We discovered the drive circuitry was exactly the same studying the board part number. I later found an article in Computer Shopper Magazine that used an RLL controller with the 3650 and gave info on the card they. I bought one, reformatted the drive and suddenly was the proud owner of a 63Meg drive. I still have this drive today.
G=C800:5 (seared into my memory). Back in the '80s, I spent a lot of time installing ST225 and ST225N (SCSI) drives. The move to RLL was mostly with Rodime equipment but there was no reference to this manufacturer in the video... maybe they were local, and not global players. Great video - thanks - looking forward to part 2.
I worked for U-Haul about 30 years ago maintaining the computers at several locations. I brought an extra hard drive to copy all there information to then I found all the bad sectors on their drive then entered them in when reformatting. Then I would copy all their info back on the newly formatted drive. I also did the backup on 5 1/4 floppies before getting an extra hard drive.
@@lucasRem-ku6eb Mainly just for maintenance and possible updates. The computers uploaded all the days transactions during the night to the servers in Phoenix.
My first hard drive was bought with a PC-XT clone, the ST-238R. 32mb RLL. I later upgraded the PC to some AT clone and for a while still kept this drive. Wing Commander II with speech (yeah i put a VGA and Sound Blaster in it) needed like 30mb so it filled the entire thing with barely some space left for MS-DOS. Sadly i don't have it anymore, but the case was identical to the ST-205 shown here. It was replaced with an IDE 80mb Seagate, which i still have, and still works (ST3096 iirc). This was my second computer, first was an Apple //c. Adrian probably knows this (hopefully): Those drives do not park their heads on their own (unlike later IDE drives), you had to use a program to park it. Fail to park it, means the heads could be left floating around, so if you take the drive out and accidentally bump it (like Adrian did) you could easily scratch the platters and damage the heads. You would normally only park the heads when you intend to physically move the drive OR the whole computer. I did have some low level formatting tools, and one useful thing you could to was change the interleave to match with your system speed and improve access times. BTW those old drives were loud, compared to modern standards.
Thanks, Adrian. Great video. Brought back tons of memories. My first HD was a Seagate ST-225. 20MB that I bought new in 1987 for $500.00! I never really understood the difference between MFM and RLL except RLLs were much more expensive. All the magazines said you wanted RLL but the wallet (and wife) said MFM. I remember entering the info such as heads, cylinders, sectors and I also remember lower-level formatting the drive. I had put it into a IBM PC/XT clone from Sanyo. AN MBC-880 that I had added and additional 128K to as well as a second half height 5.25 floppy drive. As time when on and I went through other larger MFM drives and then IDE, SATA and now M.2 SSDs, I had pretty much forgotten about the MFM/RLL specs. Thank you for clearing it all up for me and I would appreciate a video about RLLs. Thank you, again!
I have one of those Seagate ST-225 and it still works 100%, which is amazing. Adrian, I hope yours does as well. For the controller, I am using a Western Digital WD1002-WA2 card. The drive is quite noisy for today's standards, but to me it is the sound of retro.
Hey i never realized it with the IDE Systems, until now, and yes your are so right, when i see the IDE Card it is Clear. Thank you for the Informations. By the way i have an old 286 PC with a MFM 20 MB (ST225) Hartdrive like that one in the Video.
Wow, this brings back memories. In (I guess) 1985 or 86 I bought a 10mb mfm harddisk for my commodore pc 10 III. Boy o boy it would store everything and more I owned on software and “only” costed me over 800 dollars as a second hand. Now we are talking about xxTb on solid state. Lol, imagine… 10Mb….. can’t wait for part 2 Adrian, also to learn about the disk size’s.
Donaldson, is a company that specializes in industrial and automotive air and oil filters. This is the third time I have watched this interesting video. I remember that I was still using a C64, (so it would have been before I bought my Amiga,1000, around 1985 ) and I remember that a 10 meg hard drive was offered for the C64, for only $10,000. How things have changed!
Adrian, this is destined to be one of my favorite videos of yours!!! SO comprehensive of an overview and explanation of this whole topic! 4:54 back in 2015 on my first experiments with the AT&T UNIX PC & Convergent Technologies MightyFrame, I DID run nearly 20 feet of ribbon cable from machine to machine connecting physical drives and to the "Gessweinator" MFM emulators running my tests! It was quite a sight to see...I'll see if I can find pictures...
Fascinating stuff. I remember my first PC, a 286 AT clone purchased in 1988, had a 20MB MFM hard drive. I believe it was a Seagate ST-225 like the one you showed. The following year I went to a computer show (remember those?), and bought a used drive of the same type. So much room for my DOS programs!
Hey Adrian, you should mention the occasional ST-506 bearing "phase" whine phenomenon. I lost count how many times I would hear my old BBS machine's RLL drive (s) temporarily howl for like 20 seconds only to hear the sound fade away. Would make me go into cold sweats every time!
Love your video’s, i remember assembling computers with MFM and RLL HDD’s. And yes the ST225 was one of them. I also remember assembling and building 50 computers a day, and put DIP memory in all the motherboards.
Thumbs up for remembering MFM Hard drives. My first PC had one. Had a separate card for the dual 5.25" floppy disk drives my dad and I put into the PC. (Yes, I remember having an A:, B:, and C: drives.) Let's also not forget the behemoth RLL drives as well, may those door stops rest in peace.
Bringing back memories! As a field engineer, I worked on mainframes and document processors supporting the banking and automotive sector. One slow- speed doc processor utilized the ST-225's and were part of a larger field replaceable unit. Therefore, many FRU's were replaced, yet the ST-225's were fine (mostly). That said... With the help of a fabricated SCSI I/F, an Adaptec 4070, and a MacGuyverized housing, I had one of the few Amigas with a whopping amount of hd storage! I remember having to use strategically placed aluminum foil to shield my makeshift assembly. 😂 Those were the days.
The ST-225 fills me with nostalgia... I used to have quite a sizeable collection of those, I used to get them out of the skip (dumpster) at work, take them home, and they still formatted... They were awfully slow and a bit on the small size... but I think you could you could use them for a game of football (soccer) and they'd still run like a dream. OOOH! And that Western Digital card was what I plugged them all into AND I had one of the Micropolis drives too... ... ... This is amazing, my early 90s "junk" collection lives on on your workbench.
I used a few ST-225s back in the day. The first drive I bought was an ST-238, which was the same platters as the 225, but used RLL instead of MFM so it held 30meg instead of 20meg. I ran that drive for so many years. Toward the end of its life, I had to turn on the computer and let it spin the drive for 30 minutes to 'warm up' or it couldn't read the data on the inner tracks. I think the heat built up and expanded some worn components to make everything fit tighter.
I purchased a new IBM PC clone many years ago. My company needed a dependable system. So this computer lasted about 9 months before the hard drive failed. After opening the case I found a RRL controller card on a MFM hard drive. That's what crashed the hard drive and our valuable data. Fortunately a silicon valley company named Rotating Memory Service was able to recover our data. I ended up suing the company that sold us the computer and got all our money back. Never put an RRL controller on a MFM hard drive!
Issue might not have been RLL against MFM, but track to head alignment drift common to stepper motor drives. Overheated room as well as means of drive mounting could get the alignment off track, pun intended.
@@danman32 The company that we purchased the computer from advertised a Seagate 40 mg hard drive when it in fact was a 20mg Seagate MFM hard drive with a RRL controller that showed it as a 40mg drive. The silicon valley drive repair shop said the problem was the incompatible RRL controller. This was the #1 drive repair shop in Santa Clara back when when 20meg was the standard drive.
My heart still jumped when you let accidentally that one drive hit the counter. I remember the days of just looking wrong at a drive would give you a head crash. I *still* wince when I see someone carrying around a laptop running with a hard drive and walking around nonchalantly ....
Seagate and Miniscribe did put a fair amount of intelligence into their MFM drives. Seagate started doing this with the ST-225, adding built-in self test functionality and precision motion control outside of the of the controller card itself, allowing for motion optimisation and microstepping, which along with a special low-impedance stepper motor gave their drives the best seek performance of any stepper motor drive.
Man, this was my career from the late 70's to 2000. I worked at Adaptec first on SCSI to ST-506 controllers and next on the IBM PC controllers. Minor correction, the original IBM controller was created by Xebec, not WD. Our president came back from the IBM PC public announcement of the built-in controller with a full system he purchaced. He assigned 3 of us (2 hardware and 1 firmware) to duplicate the controller using our chipset with the added major requirement that we has to be format compatible so that Xebec and Adaptec controllers could be interchanged. Oh and by the way, first demo had to be in 1 month!!! We did it and sold it as the ACB-2000. One of our chip guys was working on the RLL version of our MFM ENDEC chip. I shoehorned it into the ACB-2000 and it worked the first time. This became the ACB-2070 (for the RLL 2 of 7 format. BTW, the Xebec PC controller used the basic SCSI command set ans did the IDE drives.
The rumor growing up was RLL could double your MFM disk space. The reality was it didn't always work and you needed the right combination of RLL controller and MFM drive to do it. I had an MFM ST-251 in my Packard Bell 8/12 286.
I had a full height 20MB MFM drive in the 286 PC I got from my uncle when his company was getting rid of old stuff. I was very happy when I could find another 20 MB half height drive at my high school, and could double the capacity. I remember using Norton’s Calibrate to optimize the interleave values. Good times :)
I had a full height ST506 drive that I salvaged from an old mini system that was about to be scrapped. Had 5MB total storage! I paired that with an Amiga 1000 and a Spirit HDA 506 expansion with an OMTI controller. It felt like I had something so much better than an IBM XT at the time. Later I upgraded it to an ST225 with 20MB, a half-height drive. I had just as much storage as some of the PCs I used at work at the time.
Wow! You woke up some mmemories here. I'm retired and we used to use these drives at a company I worked for early. So let me cut this down. We used to use WD 1001 through WD 1004 conrollers and were running TRS 80 Mod 4 computers running the LDOS operating system. Yikes! We also had drives with removable cartridges. We used those to do backups. I found that I could talk to the controller directly. LDOS was a pretty simple system. Using the disk architecture we had mapped I wrote a sector by sector bakup/restore utility. One day Western Digital announced the end of the WD 100X series controllers. I found replacements from RMT Systems. Long story shortened I found myself on the phone with Ralph M. Tobbleman. Ralph was the designer of the WD 100X series and took the design with him when WD let him go. When we converted to IBM PC's I wrote a utility to copy data from an LDOS drive. This made those conversions a snap. I am so glad to be away from that kind of bit twiddling.
I remember those days. I found Spinrite to be a life-saver as it could low-level format without losing data. As well as recovering data and finding those bad spots.
This reminds me of the day that I met my first "PC" hard drive. It was in 78 or 79 while at school. We heard that a company was going to be in the city to do demonstrations of a new type of hard drive for personal computers. A group of us nerd girls got in the car, ditched classes and drove 100 miles to the city. We went to a seedy motel room. We were more excited to see a new drive technology than were afraid of something sketchy (times have changed). There was a guy in a room with a briefcase filled with electronics of some sort and a full height 5.25 drive connect via a ribbon cable. He gave the talk and ran some unspecified program/demonstration to show off a 5MB hard drive that none of us could hope to afford.
That’s a great story! 😊
Hehe. It sounded like a good day out. 👍
Oh this reminds me of the day I met my first guy in a seedy hotel room. He gave the talk and ran some demonstration to show off nerd girls none of us could hope to afford.
No way nerd girls existed back then....maybe one.
damn bruh how old u is
I learned A LOT! I didn't realize a half hour passed and when Adrian mentioned a part 2 I was thinking, "Part 2, why? It hasn't been that long!" I was so wrong.
I have the feeling that part 2 will be major disappointment, its a miracle if one of those drives still work but you should not be surprised if none of them does anymore. These things could fail easily, even back in the day they are too fragile and /true/ repair requires special dustless station and parts that don't exist anymore unless you have spare drives to cannibalize from...
I thought the same. I would have been completely happy to see a longer video that combined the whole project.
@@freeculture I bet you at least the ST-225 will still work. Those drives are damn near indestructible. I have a pair from 1985 that I was able to low-level format last year with zero defects.
@@freeculture I bet you that the Micropolis has a sticky bumper or a stuck brake but will work, I bet that the ST-225 works no problem, the ST-4051 works, and the PTI drive is dead as a doornail.
Great video! 😊- I worked in a computer shop in he early 90's and remember fitting these. I remember that one of the main causes of failure was people not being bothered to 'park the heads' before they turned off the PC, thus the drive stopped with the heads resting on the drive platter. Any movement of the PC would likely result in disc damage and data corruption. New low level format, new bad sectors mapped and back to the customer.
Miss being able to low level format drives. We used to have so much control over them, wouldn't need to worry about smr vs cmr either.
I find it so fascinating to learn about the history and evolution of these legacy devices. It makes me appreciate the engineering, both mechanical and electronic, that much more. Thanks Adrian.
evolution of engineering solutions
Makes you wonder how they made the first thing in the first place.. you need something to make something else..
scammers scam scams ibm harddrive japan made nec made ibm
scammers scam scams ibm harddrive japan made nec made ibm
God. I remember having to deal with the old HDD interface cards. You had to go into debug and it was something like C800G (that is not the correct address, but it was something like that) to execute the code on the HDD interface card to do things like low level format, mark bad sectors, etc. IDE was a lifesaver and a hair saver! ^-^
G=C800:5
I remember only because I had to do it so many times, it was a format routine in rom and you had to use debug in dos to run it.
@@rogerjones8809 Yeah! And the address changed depending on who made the card. All text too, so no EZ mode. You had to actually have a clue what you were doing.
Glad to have moved past that. It was no fun at all.
G=C800:5 typically
Wow... hadn't thought of that vector in years. I had a job where all I did was integrate HDDs into XT clones.
I didn't see this but added the correct info as others had replied to in your comment above.I used to install the ST-225's into IBM XT's for Chemist shop systems in South Australia back then. Ahhh Memories
This is awesome! Really enjoyed the video and can’t wait for part 2! I’ve always been extremely fascinated by MFM hard drives and have been collecting and repairing them for quite some time now. I always use period correct MFM drives in my vintage PCs for the authentic sound, speed and feel.
I would love to see a video on RLL controllers as well!
scammers scam scams ibm harddrive japan made nec made ibm
Oh man, I miss when drives were still heavy enough to be weaponized when dropped by accident. Still miss not needing an LED to tell when files were being accessed as well. Love seeing the progression and advances in engineering.
I don't miss them because they were very fragile, heavy and fragile is not a good combination... I'm still using mechanical hard drives right now and i still treat them like handling a bomb when i have to move them for some reason, from habits of those days. I wouldn't dare do things like what Adrian did in this video and back then the things were darn expensive too! I remember some of the manuals would show/tell you could not run the drive upside down or vertically or at some odd angle for example. Yes, there was a time hard drives included printed manuals, i still have mine for the ST3144 family which includes de ST3096 i own (and i actually got the ST3144 later from somebody who didn't want it as it had some bad sectors), one is 80mb the other 120mb iirc, both have their own LED and are 3 1/2", never saw drives with one after these.
The Exos 20TB drives I have deployed are pretty damned weighty. Not like this or the Seagate Elite family, but they'll dent a dome fo-sho.
At least in those days when someone stole your computer, you just had to wait about five minutes, then walk out the door, and the first person you met outside who was out of breath, was the thief!
I'm glad you mentioned that. At a sales/repair shop in 1990ish, my tech partner is working on someones XT/8088 writing machine. (Lots of authors in this area) At one point in time (1985?) author buddy upgraded to a 20MB full-height 'Winchester' as they say.
As buddy is pulling the case top forward, angle up and off, the chassis itself angled down sharply enough to drop the behemouth drive right onto the top of his toes! It wasn't screwed in OR connected to power and data cables.
Well, this was a gruesome sight! Tech buddy squeeled like a dying baby rabbit. To make matters much more unpleasant, our tech buddy is one of those who wear open toe sandals year-round (In Canada no less). Let's just say OSHA didn't approve said sandals...
It completely shattered two of buddies toes, pierced the skin and issued a healthy flow of the red life-juice.
We sorted him out, and got him an ambulance. It is true as a Motherfuck: He never walked right again.
scammers scam scams ibm harddrive japan made nec made ibm
G=C800:5 is burned into my memory from many many (many!) years ago. Still have the ST-225 I purchased in 1987 (along with the WD controller), still works today! Amazing engineering.
And don't forget to set the interleave factor.. :-)
Hah, stay away from the Kalok :D
You forgot one type of controller card, ERLL! Back in the day, my brother and I ran an AMIGA BBS out of Tacoma, WA called "The Missing Link BBS". We originally started the system using an Amiga 500 and an 80MB SCSI drive. The Seagate 80MB SCSI drive was so expensive that it literally cost my brother 2 months pay for the hard drive! In order to expand our BBS file capacity, we converted over to my IBM XT Clone, using MFM drives that were much cheaper. We had 2 120MB MFM hard drives, but that wasn't enough. We did some research and came up with a 16 bit controller card that used ERLL! ERLL doubles your capacity, 25% more capacity than an RLL controller. With this we were able to keep the board going until 1994 when the PC graphics and software availability overshadowed the Amigas. ERLL did work with MFM drives, which is what we used and they were extremely reliable. We never once had a hard drive failure during that entire 10 year time the BBS was running.
I love the sound of these drives, definitely triggers some nostalgic memories!
Agree!
I absolutely love the spin up sound and the head initialization sequence.
i remember having to park the heads when u finished with PC lol
Run defrag, or chkdsk!!
@@kynkokytsumi1931 Good habit. Not strictly needed if you were not actually moving the computer or drive, but you never know. Thankfully the IDE drives would park on their own when losing power.
@@freeculture the main problem was head crashes when power was lost or switched off as the stepper motor can't place the heads on the "landing zone" unlike voice coil drives which as you say automatically goes to the landing zone when power is lost. On stepper motor drives the head lose the ground effect and the head then scrapes the surface of the disk. Happened to a friend when he had a power cut and lost data due to head crash.
Watching this reminds me of working in a clone shop in the early 90's. When we would get a new batch of hard drives we hadn't seen before, we would have to call the manufacturer's 800 number and request a fax with the latest head/cyl/track values before we could use them. I still remember seeing the first motherboard with an integrated IDE controller that could "auto-detect" the values. I still remember the Conner CFS420's 826 16 63 values all these years later. I must have typed that in hundreds of times while I worked there. Great video as always Adrian!
I used to do a lot of Netware, from 2.0a on. The earlier versions of the MFM controller driver only supported the geometries in the BIOS ROM. Soft configuration schemes were not supported at all. I became quite proficient at hex editing ROMs to add weird and wonderful geometries.
I started out as a student PC tech at my university in 1989. I was so glad when, in the mid '90s, IDE drives had become so ubiquitous that I could stop lugging around my "hard drive bible" ... a fat book with all of the heads/cyl/track listings for hundreds of different models of disk. IDE was a massive leap forward.
Initial IDEs still needed CHS parameters put in, very few 386s had the autodetect feature, let alone 286 boards (mine did!)
I had to re-watch your "This thing can backup and emulate MFM/ST-506 hard drives for your retro computers" video from August 2021. It is indeed, in my opinion, the most powerful tool you can have to analyze and fully evaluate (and even image) every one of those hard drives. So cool that you present so much on this subject! Keep this up, it's truly my favorite Adrian's Digital Basement topic!
Great video!
I worked in a local computer store in 1989-1990 and this brings back great memories.
Some things that might help
1. The first number in these seagate drive represents the form factor.
4 - 5.25 full height
2 - 5.25 half height
1 - 3.5
2. Full height / half height refers to the bay size. So the lager 3.5 drive would still be considered half height.
3. The following numbers represents the un-formatted capacity.
St225 - 25 meg un-formatted /21 formatted
St251 51 meg un-formatted / 42 formatted
3. To setup 8 bit controllers
Run: debug G=C800:5
I still remember this 34 years later.
When I was a teenager working for a large distributor of Seagate, WD and Miniscribe, we did extensive testing for weeks continuously in a CRC bad sector tester. The machine we were using was the very same one that prints the bad sectors on top of the drive. We failed to find any MFM drives that was made less reliable by running on an RLL controller. After these tests, I never ran a drive with MFM scheme again for my own use at least. The increase in capacity was almost universally experienced as 1.5x
I think it just had to do with warrenty to cover the manufacturer's backs, I never understood how a RLL controller would stress a MFM HDD more than it should. The way I understand how it works is that RLL encodes data more efficiently than MFM does, nothing less nothing more.
I can add that back in that time, it was also common to add such drives to Amiga (with famous ALF toolkit). And everyone I knew (including me) bought a MFM drive and used it with RLL
@@Alexis_du_60 We did prove that in the sample sizes we had. Over the course of one summer we probably put 50 drives of different MFM generations under those conditions. A couple failed, but that was well within the failure rate at the time, regardless of encoding method.
@@videowatcherdeluxe It is as reliable as drives were at the time. In the early 8 or 16-bit days, I never lost a drive. Now, of course, even though drive technology is basically magic, I have lost a few, but only over the decades since.
scammers scam scams ibm harddrive japan made nec made ibm
Great video.
MFM drives were before my time as I started with IDE, so great to learn more about them.
I can remember the moment I reconfigured my early 90s hard drive controller from MFM to RLL.
My 40MB drive was transformed into a 63MB drive, just like magic!
I can't recall the brand of drive, but it seemed to work perfectly fine with RLL encoding.
The standard IBM PC interface used IRQ5 for these original MFM controllers; but when Tandy produced their PC Clone Tandy 1000, they deliberately "proprietized" the interface and changed the interrupt. Standard, off-the-shelf controllers wouldn't work despite bus-level compatibility; until I saw a magazine article with the fix; take a screwdriver and scratch out a trace on the controller card. That changed the default interrupt the board used to one the Tandy could talk to. And, yes, it worked 😁. It was only when I took my 1000 to a Tandy repair center for a warranty video chip change out that a store manager wailed at me for having a "non-Radio Shack hard drive," and I told him, "yeah, and it has nothing to do with the character generator ROM issue." He hushed, changed out the ROM, and I left happy 😁
GAWD, talk about bringing back memories. I've installed countless ST-225 drives into IBM PC's and XT's back in the day. Even remember using DEBUG to jump into the ROM address on the controller card which the command was G=C800:5 or telling Debug to GO to address C800 Byte 5 to start the routine in ROM on the card, WHICH then brings up the menu and you go from there.
Keep up the good work there Adrian. love watching your vids. as I LIVED in the era of Tandy TRS-80's (I had one) and yeah became a digital engineer.
I think the reason that RLL was never really used on floppy drives is because, as you mentioned, RLL can be difficult to decode properly if you don't have an extremely consistent disk rotation speed, and floppy disks can vary a _lot_ more in rotation speed compared to hard drives. If the exact conditions of the disk itself change at all over time, an RLL signal written to the disk could become basically unreadable, and with floppy disks that could easily happen due to a little bit of dirt, or a slight stretching of the (plastic) magnetic media, or just temperature changes, etc.
Yes....... and that's also why IOMega Bernoulli could achieve its data density.
The early on the MFM ST225 was a 3600 RPM stability failed version of the RLL ST238 . The early RLL controllers could not RPM synchronize well with drives that did not meet stringent requirements .
A completely separate company modified on of these modulation and controlling mechanisms to make things called video image files where each track of the multi platter had stored on it half of single NTSC monochrome video frames using a pulse width modulation technique . The ST225 quality version would create NTSC image flagging whereas the ST238 quality version had way less flagging .
A likely reason why ST225 were more prevalent is because the ST225 was easier to get information out of as the spindle would ware out , which without which a later generation controllers could compensate for would send a ST238 to the scrap heap .
@@donwald3436 I dunno about that.. I can't find any real details on the media encoding used on those sorts of devices, but just using RLL would not account for the large capacity improvements by itself, and they also used flexible (plastic) media in non-sealed environment which would have similar stability problems, IMHO, so I wouldn't be at all surprised if they still used MFM or something like it too, for the same reasons.
The main reason for the higher capacities in those drives, and successors like the Zip drive, etc is, I'm pretty sure, a combination of higher coercivity media and high write currents allowing for higher magnetic field density, along with more precise head positioning allowing for many many more tracks on the disk (poor track density due to imprecise head positioning was really the largest source of "wasted capacity" in most traditional floppy designs).
Though it is possible that the LS-120 "SuperDisk", which had an optically-coded reference track may have had precise enough real-time timing information to make RLL also be feasible for that. I can't find anything that actually says one way or another, though...
@@foogod4237 I remember (hearsay!) that some folks on the FIDO hardware groups mentioned that not all MFM HDDs were able to be used with RLL. Specifically the ST-4096 seems to have had overheating issues with the driver electronics.
@@securityteam Electronics problems wouldn't really make any sense, IMHO. There's no difference to the drive in the sorts of raw signals it's being asked to work with. The only difference between MFM and RLL is how the raw on-off patterns are translated to data bits by the controller. The signal timing and frequencies are all essentially the same from the drive's perspective.
There were some MFM drives which couldn't reliably be used for RLL because their spin rates were not reliable enough, so the controller sometimes had a hard time reading back what it had written because the bit timings varied too much from one part to another (precise timing is much more critical to correctly interpret RLL codes).. As AI T mentioned, that was the main difference between the ST-225 (MFM) drives and the ST-238 (RLL) drives.. They were basically exactly the same drive, it's just that one was tested and certified to be guaranteed to have a rotation speed which was stable enough for RLL, while the other wasn't guaranteed (but many people did try to do RLL with ST-225s anyway, and it did sometimes work).
Brilliant Adrian - I have fond memories of these type of drives, particularly the ST225 - my very first PC clone had one of those! Very informative as usual and I've learnt some more about these things!
Had fond memories of this when I was young and had my 286 with a 35MB MFM HDD, nice to see and learn finally how they worked, currently my oldest system is a 386 DX40 hope I come across a working 286 or XT still with One of these. Thanks for the new knowledge, i must say you're the best in explaining the concepts and theories behing these old tech. Thnx again.
Yes, I definitely want a video on RLL drives!
Adrian is such a tease! Ofc we want an RLL video!
And a video about ESDI.
Had a XT clone with a full height 5¼" 650MB Micropolis and a half height 350MB Maxtor. Almost 1GB of disk space and MS DOS 3.3...
Adrian - this was a ton of fun - I haven't heard these acronyms in years, thanks for the drive interface rewind. I had forgotten that ATA was basically ISA on a ribbon cable. So many alphabet soup acronyms I used to see every day booting my computer... ATAPI, IDE... so many acronyms browsing PC advertisements in the 90's... Ultra DMA, Ultra ATA...
OH and then remembering the C/H/S... I think Type 47 was the user definable one.
Excellent presentation. Thank you. I remember working with all of these back in the late 80's.
My ST-506 hard disk was connected to an Z80 computer using a XEBEC 1410 controller around 1982. I was working on a Z8000 based system at the time using the ST-506 and XEBEC 1410 controller so just ported one over for my 256k CPM 3.0 system. Was equivalent to 5 double density / double sided 80 track floppy disks - WOW!
I later used the Adaptec RLL ISA bus controller to update my MFM drives for more storage. Never had a problem with the RLL encoding.
I remember the first MFM hard drive I installed on my IBM.
It certainly was a learning experience. ( no internet yet and BBS's would take a few more years to mature )
From memory, the hard drives bought after the 80's also had the head and sector info along with the bad sectors printed on the top.
My bbs ran a 20 c drive and a 40mb d drive...mfm. ...great times
IBM sold them in systems before April 1981.
removed them all for SCSI and IDE, on 8 bit ISA interface cards.
Why he needs borken old disks ...?
Yeah the bad block map - it was rare to see blank ones.
@@the_kombinator I remeber spending time looking for the smallest list on the stack...
Sorry for laughing at the "finger acrobatics" 😁. No one saw me. Thanks Adrian for yet another very interesting video. I still remember my 3.5" HDD, possibly about 200MB, "mounted" in a plastic ice cream box outside my Amiga 1200. I had to drive the computer and harddrive with a rebuilt AT-PSU. The 1200 was nice by giving us an internal IDE-connector, compared to the 500 where one had to use the left side expansion port for such pleasures.
Brilliant, Mr Black. I do love thèse walks on memory lane. Especially because I wasn’t a pc lover at that time so I’m on full discovery mode. But it helps a lot to understand why we are where we are now. So, please, keep them coming, RLE included 🙏
The sound sooo good, first drive i ever heard...
I enjoyed your post on MFM hard drives. I am an older adoptee of computers, not until 1988 at the age of 37 did I acquire my first a Commodore 64. In the early '90s I stated rebuilding and selling used computers having self taught myself to do so. Most st first had 10 or 20 meg MFM hard drives with most being full height drives at first. Later on the hard drives were 20 or 40 megs in various configurations up to 80 meg total. I usually left the first drive uncompressed and used drvspace or dblspace on the second drive to get more storage space. I used a lot of different brands and capacity MFM drives the largest being the old full height Maxtor and Miniscribe 80 mb drives, the top covers being cast with the company name in large raised letters on top..A work of art and not cheap either. One thing I did was I tried to buy the hard drives with the least number of errors. Also hard drive setup was time consuming sometimes taking hours on a very large drive. Another thing I did for some of my own drives was to format n MFM drive with a RRL controller and then drvspace the drive. On a 40 meg drive I would end up with somewhere around 90 megs of space with the conversion losses causing you to lose theoretical space. It made for a slow drive with lots of clicking and drive activity light going nuts but it worked.
8:36 I suspect you mean "binary encoded" not "binary coded decimal" for the head select signals. BCD would require 5 bits to get to 16 heads. :) Really cool video, and very excited for the next installment to see which ones will work. I know I have my predictions!
TIL Shugart became Seagate. That's one of the things I love about episodes like this one! At a previous job, one of the other techs told me of some trouble he had when he was working on a notoriously sensitive test fixture belonging to one of our customers (if it ever moved on the floor even a quarter inch, a factory tech had to come out and recalibrate the fixture). It was driven by an ancient IBM PC with full-height hard drive, and despite numerous requests, the company saw no need to upgrade the PC with a newer model. That changed one day when the tech was downloading test results onto a 5 1/4" floppy (this was in the late '90s). He suddenly heard a CLANG so loud that he thought something had fallen off the PC shelf onto the floor! The company (reluctantly) sent out a replacement PC, and you could tell that although it finally had a 3 1/2" drive, it was literally scraped together from whatever spare parts they had laying around!
I actually have on old ST-4051 on my shelf that I've had issues pulling data off of! Very interested to see your approach. And you should definitely make a video on RLL! A lot of pre-IDE disk technology doesn't get a lot of attention and it could help people that run across these older drives so they don't accidentally get written off as dead.
Yes! I am part of that demographic!
One of the first PCs I had is one that I built from scraps at an electronics recycling shop in the early 1990s. Of course nothing was properly labeled, and there were a lot of junk parts - but I learned a bunch about hardware. I ended up buying an RLL controller assuming it was just a regular MFM (same connectors means it's the same thing, after all -- right? Ha!) and was super surprised to find my 10MB ST-412 formatted out to 15MB! I used RLL controllers on all kinds of "MFM" drives from that point on, and never had an issue. It wasn't until much later I learned that using an MFM drive with an RLL controller "wouldn't work right." Ah, those were the days... :D
Thanks for keeping all this history alive.
But, it actually does work fine.
@@mikebarushok5361 Most of the time.
I was waiting for a decent video regarding MFM/ST506 and just joined as patreon! I couldn't be luckier!
I installed an ST225 with an Amstrad PC-1640 which I reconfigured to work with a Sony TV. The hard disk was a little noisy but worked very well with an MFM interface card. I installed it outside the case so I could retain the twin floppy drives, for some strange reason that I can't remember.
I didn't start to work on PC's until I was a young adult in the mid '90s. So while I saw these old MFM hard drives back in the day I never really learned about them. So I find your video contact about this subject fascinating and educational.
Thanks for this video. I remember these drives from the beginning of my career. By the time I planned the purchase of my first PC (in 1989), I was trying to make sense of the specifications for MFM, RLL and ESDI hard drives. The decision ended up being made for me by the system builder, who put in a 90MB SCSI hard drive that worked fine once the computer finished booting, but that old 8-bit SCSI controller took 3+ minutes start up. The SCSI card and hard drive were replaced in 1991 or '92 with an IDE interface and hard drive. (That computer was literally a bargain basement build, and in retrospect, I think it was pieced together as a conglomeration of questionable purchases previously made by the builder. It was built on an ELT-200B 80386DX motherboard that used hard-to-find, expensive SIPP memory modules and proved to have a lot of strange compatibility issues with all sorts of components and peripherals.)
I would love to see a video about RLL hard drives, and also ESDI. It seems to me that these were two technologies that were only around for a short time and didn't make a lot of splash in the market, although I know ESDI was better known because IBM used it in at least some of their PS/2s.
This video hits FAST and HARD with the avalanche of knowledge of so many things I've wondered about and were all answered. I feel like it's in your very top tier quality videos. Kudos. Also, I had a ST-225 drive inside an external Macintosh Hyperdrive FX/20 enclosure and it failed while I was changing diapers on my then baby son midpoint into copying critical data on diskettes. Thankfully I found the files in a Classic 1 later.
Coming from an Amiga 500, I didn't own a computer with a hard drive until well into the 3.5" era, ( 386 PC). So I skipped over all the earlier shenanigans. This video was really informative. 👍
I used three MFM-drives on my Amiga 2000. Two of them were outside the computer and I swaped between them. The one in the Amiga had Workbench. Used one drive for games and the third to utilities.
Great video! My first PC was an AT&T PC 6300 with a 20mb RLL encoded hard drive. The only thing I have left of it is the 5.25" floppy drive. I miss it. Looking forward to part 2 and some Coleco Adam content.
I too missed having an AT&T PC 6300..... so I hunted a nice clean one down and it's safe and sound on a shelf in the basement. I even managed to track down the matching color monitor so now I have the option of using monochrome or color on rainy days when I fire it up for some nostalgia. Mine came with a trusty ST-225 factory installed hard drive in it, some day I'd like to track down an original Olivetti hard drive for it too.
@@RogerWilco486 that's awesome! I wish I never got rid of mine!
Besides being a trip down memory lane, this was an eye opener. Many of the early stuff I was learning had no one to lay it out like you just did. AWESOME JOB!! thanks .... signed, 1980's engineering student.
The memories of putting my first 5mb full height hard drive in my home soldered XT. "Debug G=C800:0005" is the first thing that comes to mind. Being able to use PCTools and Turbo Pascal without switching floppy's was a whole new experience.
My first harddrive was an indeed a 20MB Seagate ST-225 MFM drive - but I used it for my Amiga 500.
This was actually a low-cost solution at the time, using a PC-MFM controller card made by OMTI and a simple adaptor card (two sockets for the Amiga and the OMTI and a few logic chips). This adaptor was built from a magazine article at the time and it worked like a charm.
I remember its performance being about 300 KB/s which was enormous at the time. Many programs weren't larger than that so loading times were quite snappy.
The only real downside being that one couldn't throw games willy nilly on it as they either used copy-protected floppy disks or were simply not behaving very well (they used all RAM, thrashed the OS or used their own file system, IIRC). At least most of the games I had.
For example, I remember being super annoyed that I still had to swap floppies when playing Ultima V.
When I saw the screenshots of Ultima VI on PC in its 256 color VGA glory I sold the Amiga immediately for a good price (I had put the 500 mainboard into a PC case, the keyboard in a separate shell with a spiral cable and this contraption used a regular PC power supply with that big clunky switch). That was in 1991 and since them I'm happily in the PC camp.
I love hearing from folks whose first computer was not an IBM. Actually I started on IBM mainframes before PC's even came out. My first desktop computer for work was TRS-80 Mod 4. Yep, we put hard drives in them.
@@crankshaft3612 Oh the Amiga wasn't my first computer but rather the fourth...
(ZX81 -> Atari 400 -> Atari 130XE -> Amiga 500 -> Lots of PCs made from components...)
Loved the introduction. I've seen and used all these types over the years but didn't put together their genetic lineage and evolutionary family tree before. Fascinating!
Back in the late 1980's I was the guy if you needed your seagate drive repaired, or at least have the data recovered. I worked on other drives as well. We had no clean room, and really when these drives were bad, a cleanroom was the least of your problems. I had one st225 that had a head crash that was allowed to run for a coup[le weeks, so the first step was to blow out the ounce of aluminum dust from the drive, then to carefully wash it completely clean before attempting the recovery. In that case I was able to recover the one file the user was asking for, after I showed that I was not going to be able to pull a full recovery because one platter was scrubbed free of media. With another drive recovery for AMD, they had a chip design on the drive, and no backup. AMD got lucky, I was able to perform a repair of the drive's electronics, and pull their data over onto a better drive, and recover their chip design. It was a rll drive, which were the worst of the drives because rll drives tended to loose data long before the drive "went bad". RLL drives used the SAME controller as the MFM, so I don't know where you are getting that, you are completely wrong on that... However there were many different oddball controllers, so perhaps you are talking about one of those... But the standard western digital controllers for both the xt and at would transparently operate both mfm and rll drives. AMD got lucky, I was able to perform a repair of the drive's electronics, and pull their data over onto a better drive, and recover their chip design.
I stopped doing drive repair when the IDE drives came out as the "smart" drives wouldn't even spin up if they thought they were dead, so there was no mode I could run them in to do the sector reads I needed to do, and the servo media readers were not yet available.
Interestingly, I had a 100% success rate with the MFM drives, 60% success rate for rll... that went to 30% when IDE came out. It was an easy decision not to work on IDE's. I had some success, but it wasn't enough to justify the labor, as I couldn't charge for a failed repair, and raising my rates 300% to make up the difference didn't sit well.
" RLL drives used the SAME controller as the MFM" Yes, thanks for confirming my memories of that. Perhaps the drives I used (can count on one hand) were those Western Digital ones, because I remember MFM and MFM/RLL being exactly the same hardware, I was under the impression that "RLL" was really just a format.
Both so-called MFM and so-called RLL disks (or rather controllers) are using MFM modulation and RLL encoding. The so-called MFM disks are using RLL(1,3) encoding, while the so-called RLL ones are using RLL(2,7) encoding.
Can't wait for part 2!! I'm weird and used to collect old HDD as a kid. Had quite the collection of these beasts! Luckily I still have a stack. Something about hearing them spin up, low level formatting... it's relaxing to me. I have recently found a pair of Shugart 1004 8" HDD that I would LOVE to get going someday! Just no idea how to 😅
Adrian, your material is interesting, informative and of top quality! Eagerly awaiting part II of this series
My first HDD was a 20 MB MFM drive that was dropped while left unparked and was discarded as completely unusable/dead. I was working in the IT dept at a government office as a student (summer job). I was able to get the drive for free (one man's trash is another man's treasure). I got myself a used MFM controller card (still expensive at the time), and spent countless hours running tools on the drive to mark blocks as bad. I think about 10-15% of the drive was unusable, maybe more, but it survived and I used it in my 1st PC for years as I went through school. Then my sister used the PC for a few more years until the motherboard died. The HDD was still cruising. This thing was a tank.
I had a bit of experience swapping controller cards back in 1988. At that time I bought a Miniscribe 3650 which is a 40Meg drive and 8 bit MFM controller using it for several months. Later a friend of mine bought a Miniscribe 3675, a 60Meg drive with an RLL controller. We discovered the drive circuitry was exactly the same studying the board part number. I later found an article in Computer Shopper Magazine that used an RLL controller with the 3650 and gave info on the card they. I bought one, reformatted the drive and suddenly was the proud owner of a 63Meg drive. I still have this drive today.
As always, your content is very enjoyable.
You're one of my favourite content creators that fill a nerdy interest that TV could never fill for me.
Great video Adrian, I'll wait for part 2 before I attempt to repair my bad ST-225s
G=C800:5 (seared into my memory). Back in the '80s, I spent a lot of time installing ST225 and ST225N (SCSI) drives. The move to RLL was mostly with Rodime equipment but there was no reference to this manufacturer in the video... maybe they were local, and not global players. Great video - thanks - looking forward to part 2.
Very blast from the past stuff. Thanks for the great content.
I worked for U-Haul about 30 years ago maintaining the computers at several locations. I brought an extra hard drive to copy all there information to then I found all the bad sectors on their drive then entered them in when reformatting. Then I would copy all their info back on the newly formatted drive. I also did the backup on 5 1/4 floppies before getting an extra hard drive.
What data U-Haul was collection, constummers, destinations etc.
Not so many data i guess, you ever did a restore Not needed i guess, Tax people only ?
@@lucasRem-ku6eb Mainly just for maintenance and possible updates. The computers uploaded all the days transactions during the night to the servers in Phoenix.
My first hard drive was bought with a PC-XT clone, the ST-238R. 32mb RLL. I later upgraded the PC to some AT clone and for a while still kept this drive. Wing Commander II with speech (yeah i put a VGA and Sound Blaster in it) needed like 30mb so it filled the entire thing with barely some space left for MS-DOS. Sadly i don't have it anymore, but the case was identical to the ST-205 shown here. It was replaced with an IDE 80mb Seagate, which i still have, and still works (ST3096 iirc). This was my second computer, first was an Apple //c. Adrian probably knows this (hopefully): Those drives do not park their heads on their own (unlike later IDE drives), you had to use a program to park it. Fail to park it, means the heads could be left floating around, so if you take the drive out and accidentally bump it (like Adrian did) you could easily scratch the platters and damage the heads. You would normally only park the heads when you intend to physically move the drive OR the whole computer. I did have some low level formatting tools, and one useful thing you could to was change the interleave to match with your system speed and improve access times. BTW those old drives were loud, compared to modern standards.
Incredible video ! Thank you for taking the time to create this
Collectors of these things like me really love old hard drives!! Great video man!
Thanks, Adrian. Great video. Brought back tons of memories. My first HD was a Seagate ST-225. 20MB that I bought new in 1987 for $500.00! I never really understood the difference between MFM and RLL except RLLs were much more expensive. All the magazines said you wanted RLL but the wallet (and wife) said MFM. I remember entering the info such as heads, cylinders, sectors and I also remember lower-level formatting the drive. I had put it into a IBM PC/XT clone from Sanyo. AN MBC-880 that I had added and additional 128K to as well as a second half height 5.25 floppy drive. As time when on and I went through other larger MFM drives and then IDE, SATA and now M.2 SSDs, I had pretty much forgotten about the MFM/RLL specs. Thank you for clearing it all up for me and I would appreciate a video about RLLs. Thank you, again!
This has been cool. Looking forward to part 2. It was fun to see the five rails. I haven’t thought about those in many many moons.
Fascinating! Never realized that the original IDE was a direct connection to the ISA bus!!!!!
I have one of those Seagate ST-225 and it still works 100%, which is amazing. Adrian, I hope yours does as well. For the controller, I am using a Western Digital WD1002-WA2 card. The drive is quite noisy for today's standards, but to me it is the sound of retro.
I recorded this months ago, but I am just uploading it - ua-cam.com/users/shortsR8aNDe4BpVg
As I remember that card supports 2 drives.
@@crankshaft3612 it does, and a floppy drive as well. That card is super long, like a VLB video card.
I love this history! Really interesting, thanks!
Incredibly informative, nice job as always! 😃
Hey i never realized it with the IDE Systems, until now, and yes your are so right, when i see the IDE Card it is Clear. Thank you for the Informations. By the way i have an old 286 PC with a MFM 20 MB (ST225) Hartdrive like that one in the Video.
Wow, this brings back memories. In (I guess) 1985 or 86 I bought a 10mb mfm harddisk for my commodore pc 10 III. Boy o boy it would store everything and more I owned on software and “only” costed me over 800 dollars as a second hand. Now we are talking about xxTb on solid state. Lol, imagine… 10Mb….. can’t wait for part 2 Adrian, also to learn about the disk size’s.
Now that brings back some memories, because I had exactly that WD1006 controller with that Micropolis 1325 in my first PC. Great episode!
Donaldson, is a company that specializes in industrial and automotive air and oil filters. This is the third time I have watched this interesting video. I remember that I was still using a C64, (so it would have been before I bought my Amiga,1000, around 1985 ) and I remember that a 10 meg hard drive was offered for the C64, for only $10,000. How things have changed!
Great video! I learned a lot about hard drives that I didn't know before. Thanks!
Adrian, this is destined to be one of my favorite videos of yours!!! SO comprehensive of an overview and explanation of this whole topic! 4:54 back in 2015 on my first experiments with the AT&T UNIX PC & Convergent Technologies MightyFrame, I DID run nearly 20 feet of ribbon cable from machine to machine connecting physical drives and to the "Gessweinator" MFM emulators running my tests! It was quite a sight to see...I'll see if I can find pictures...
I like a good amount of "ado" as a appetizer, before the main course of repair and troubleshooting
Fascinating stuff. I remember my first PC, a 286 AT clone purchased in 1988, had a 20MB MFM hard drive. I believe it was a Seagate ST-225 like the one you showed. The following year I went to a computer show (remember those?), and bought a used drive of the same type. So much room for my DOS programs!
Hey Adrian, you should mention the occasional ST-506 bearing "phase" whine phenomenon. I lost count how many times I would hear my old BBS machine's RLL drive (s) temporarily howl for like 20 seconds only to hear the sound fade away. Would make me go into cold sweats every time!
I used to lubricate the bearing earth clip by scribbling with a soft pencil - five minute job and customers thought I was a miracle worker!
I love this video !! Looking forward to in-depth diagnosis in the following parts
Love your video’s, i remember assembling computers with MFM and RLL HDD’s. And yes the ST225 was one of them. I also remember assembling and building 50 computers a day, and put DIP memory in all the motherboards.
Oh thank god for you! I was born in 88 and never really used these. I need to recover data for someone and never used pre IDE drives
Loved the in-depth background explanation, thanks
watching and listening to Adrian's videos is like eating comfort food
Had a Miniscribe 3650 in the family's 1984 vintage IBM PC...picked it over an ST-238R...what a noise maker! Never failed us, though...
Thumbs up for remembering MFM Hard drives. My first PC had one. Had a separate card for the dual 5.25" floppy disk drives my dad and I put into the PC. (Yes, I remember having an A:, B:, and C: drives.) Let's also not forget the behemoth RLL drives as well, may those door stops rest in peace.
Bringing back memories!
As a field engineer, I worked on mainframes and document processors supporting the banking and automotive sector. One slow- speed doc processor utilized the ST-225's and were part of a larger field replaceable unit. Therefore, many FRU's were replaced, yet the ST-225's were fine (mostly). That said...
With the help of a fabricated SCSI I/F, an Adaptec 4070, and a MacGuyverized housing, I had one of the few Amigas with a whopping amount of hd storage! I remember having to use strategically placed aluminum foil to shield my makeshift assembly. 😂 Those were the days.
The ST-225 fills me with nostalgia... I used to have quite a sizeable collection of those, I used to get them out of the skip (dumpster) at work, take them home, and they still formatted... They were awfully slow and a bit on the small size... but I think you could you could use them for a game of football (soccer) and they'd still run like a dream. OOOH! And that Western Digital card was what I plugged them all into AND I had one of the Micropolis drives too... ... ... This is amazing, my early 90s "junk" collection lives on on your workbench.
I used a few ST-225s back in the day. The first drive I bought was an ST-238, which was the same platters as the 225, but used RLL instead of MFM so it held 30meg instead of 20meg. I ran that drive for so many years. Toward the end of its life, I had to turn on the computer and let it spin the drive for 30 minutes to 'warm up' or it couldn't read the data on the inner tracks. I think the heat built up and expanded some worn components to make everything fit tighter.
Your thought about the RLL experiment got me very interested in seeing you do that. Either way a segment on RLL would be cool!
Keep this kind of content Adrian! is ver intresting. Greetings From Argentina!
I purchased a new IBM PC clone many years ago. My company needed a dependable system. So this computer lasted about 9 months before the hard drive failed. After opening the case I found a RRL controller card on a MFM hard drive. That's what crashed the hard drive and our valuable data. Fortunately a silicon valley company named Rotating Memory Service was able to recover our data. I ended up suing the company that sold us the computer and got all our money back. Never put an RRL controller on a MFM hard drive!
Issue might not have been RLL against MFM, but track to head alignment drift common to stepper motor drives.
Overheated room as well as means of drive mounting could get the alignment off track, pun intended.
@@danman32 The company that we purchased the computer from advertised a Seagate 40 mg hard drive when it in fact was a 20mg Seagate MFM hard drive with a RRL controller that showed it as a 40mg drive. The silicon valley drive repair shop said the problem was the incompatible RRL controller. This was the #1 drive repair shop in Santa Clara back when when 20meg was the standard drive.
My heart still jumped when you let accidentally that one drive hit the counter. I remember the days of just looking wrong at a drive would give you a head crash. I *still* wince when I see someone carrying around a laptop running with a hard drive and walking around nonchalantly ....
Seagate and Miniscribe did put a fair amount of intelligence into their MFM drives. Seagate started doing this with the ST-225, adding built-in self test functionality and precision motion control outside of the of the controller card itself, allowing for motion optimisation and microstepping, which along with a special low-impedance stepper motor gave their drives the best seek performance of any stepper motor drive.
Man, this was my career from the late 70's to 2000. I worked at Adaptec first on SCSI to ST-506 controllers and next on the IBM PC controllers. Minor correction, the original IBM controller was created by Xebec, not WD. Our president came back from the IBM PC public announcement of the built-in controller with a full system he purchaced. He assigned 3 of us (2 hardware and 1 firmware) to duplicate the controller using our chipset with the added major requirement that we has to be format compatible so that Xebec and Adaptec controllers could be interchanged. Oh and by the way, first demo had to be in 1 month!!! We did it and sold it as the ACB-2000. One of our chip guys was working on the RLL version of our MFM ENDEC chip. I shoehorned it into the ACB-2000 and it worked the first time. This became the ACB-2070 (for the RLL 2 of 7 format.
BTW, the Xebec PC controller used the basic SCSI command set ans did the IDE drives.
The rumor growing up was RLL could double your MFM disk space. The reality was it didn't always work and you needed the right combination of RLL controller and MFM drive to do it. I had an MFM ST-251 in my Packard Bell 8/12 286.
I had a full height 20MB MFM drive in the 286 PC I got from my uncle when his company was getting rid of old stuff. I was very happy when I could find another 20 MB half height drive at my high school, and could double the capacity. I remember using Norton’s Calibrate to optimize the interleave values. Good times :)
I had a full height ST506 drive that I salvaged from an old mini system that was about to be scrapped. Had 5MB total storage! I paired that with an Amiga 1000 and a Spirit HDA 506 expansion with an OMTI controller. It felt like I had something so much better than an IBM XT at the time. Later I upgraded it to an ST225 with 20MB, a half-height drive. I had just as much storage as some of the PCs I used at work at the time.
Wow! You woke up some mmemories here. I'm retired and we used to use these drives at a company I worked for early. So let me cut this down.
We used to use WD 1001 through WD 1004 conrollers and were running TRS 80 Mod 4 computers running the LDOS operating system. Yikes! We also had drives with removable cartridges. We used those to do backups. I found that I could talk to the controller directly. LDOS was a pretty simple system. Using the disk architecture we had mapped I wrote a sector by sector bakup/restore utility.
One day Western Digital announced the end of the WD 100X series controllers. I found replacements from RMT Systems. Long story shortened I found myself on the phone with Ralph M. Tobbleman. Ralph was the designer of the WD 100X series and took the design with him when WD let him go.
When we converted to IBM PC's I wrote a utility to copy data from an LDOS drive. This made those conversions a snap. I am so glad to be away from that kind of bit twiddling.
I always wondered how MFM hard drives were configured. Thanks!
Can't wait for this series... I had a few of those back in the early 2000's but never got around to firing them up....
I remember those days. I found Spinrite to be a life-saver as it could low-level format without losing data. As well as recovering data and finding those bad spots.
Steve Gibson to the rescue.