@@jurajlutter yep, I ment the entire 2BSD family, so 2.8 or 2.10 or even earlier would be an option, but obviously, I'd also run 2.11BSD (which I do on my PiDP-11 :)
some where some old guy got mystery wood when that egg dropped....it would be evdence of a just god if he got know it's been found and seen by the whole HELLORD world
Hey there... don't give up on the drive with the alignment optic thing quite so fast.... I don't think those drives have any permanent index data on them, if everything is working, you might well find the a simple low level format will fix it... I did this once back in about 1994 for a bet, and actually removed the heads and all the platters from a drive, literally cleaned them with window cleaner spray, put the thing back together and got it fully working again after a low level format... where, in fact, no bad sectors were found,
That reminds me how I revived a Miniscribe 10 MB hard drive which had internal contamination. I cleaned the platter with Windex or something like that, I gently cleaned the heads with IPA, did a low-level format and the drive was good to go with just a few KB of bad sectors. That's a good thing that old MFM drives are so stupid, as if they're not too much damaged they can be brought back to life with relative ease and without special tools. Fun fact: the platter was not coated with the iron oxide typical of old 5.25 MFM drives, but was already an anodized one. That's quite odd for a drive so old.
I kind of remember the low level format being in the ROM of the drive and there was some way you just ran it. That's going back to the first drive I purchased in 1987 so might be a little off there.
Probably, but then you just erased whatever was on it, which might be interesting. I'd intercept the alignment sensor to see what it sees and then manipulate the signal to get back into alignment. Oscilloscope and a poti should be enough for that.
@@JimHarrigan On later IDE drives there was critical structural data, firmware for the drives microcontrollers etc stored on the drive itself, so you could not low level format it... but older drives were very basic.
@@UsagiElectricI don’t know why, but whenever I need to open up old electronics, those “warranty void” stickers always make me hesitate, even though the warranties expired decades ago.🤪
@UsagiElectric with the exception of things like phones (where you can't do anything once you're in), I always slice the stickers. even if I'm not opening the thing up, stickers are destroy-on-sight.
About 30 years ago when I was a teenager I took the cover off of a working 20MB MFM drive in an XT (they were worthless old junk at the time) and powered it up so that I could watch the heads move. It was fascinating but the drive started developing a bunch of bad sectors after only a few minutes of operating that way. I think that running a drive with the cover off disturbs the airflow and causes the heads to crash even if there is not a bunch of dust in the air so I would never power up an open drive I had any intention of trying to fix. I remember disassembling several other 10-20MB drives back then because they weren't worth anything, really wish I'd saved all that stuff but that's how it goes. Old junk becomes desirable collectibles but you never know what will be desireable.
The problem is microscopic dust -- what always floats around that you can't see is nonetheless bigger than the gap between the platter and the floating head. Some dust lands on the platter and sticks, and the head hits it and bounces -- microscopically, but still enough to screw up the read and eventually damage it the surface, and the head. That's why data recovery is done in a clean room (if they need to open up the drive), with literally NO dust, microscopic or otherwise. I once scrapped some ancient nonworking RLL drive, only to shortly learn that in the recovery market, the durn thing was worth (are you sitting down??) .... $900 !!! That's Nine Hundred Dollars, not a typo. ARGH!
Even knowing those drives were probably cooked, watching them be opened was difficult. 😅 It took me YEARS to learn to be OK with opening drives, even when known to be dead. But then I found the magnets, and all was good. Also, as a fellow auto enthusiast, I have far too much sympathy for the 10 minute job that takes 2 days. 👍
And some of the platters made beautiful wind chimes. But yes, it was painful watching them opened, and knowing that now the platters and heads will have dust damage. They _might_ have large enough sector tolerances to still read, but for how long?
@@Reziac I wish I lived in an area I could turn some of my superfluous platters into wind chimes. Unfortunately, in the Appalachian region of Virginia, we either have no wind, or tropical storm force. (Already lost 2 wind chimes to the latter.) Still thinking about different crafty ideas for them. I like the iridescent coatings some of the platters use. Seems like a light sculpture opportunity of some kind.
That "bearing" noise actually sounds like noise the static eliminator makes on these old drives. The static eliminator strip is located outside the sealed compartment on the bottom side of the drive. Some times it's dirt and wear that makes them squeal. You can try using a little IPA on a piece of notebook paper dragged between the spindle and static eliminator. If not dirty, try slightly increasing the tension on the static eliminator. Well, I wrote this before watching the entire video. Glad you found it. I've heard that noise way too many times in days gone past.😅
And type the command 'history' to see a full list of remembered commands. If you type your password into an open prompt by accident, you can use the same command to clear individual entries.
Thanks! Though I did already know that shortcut. When I'm doing screen capture for a video I always retype the full command so that it gives the viewer, especially those not familiar with Linux, time to visually locate the command and mentally parse what is being typed before it flies off the screen.
It's called the "Light Positioning System". These drives have a servo surface but it is actually possible to configure them to ignore the servo surface and just use the LPS for tracking.
I noticed a 'void warranty' sticker on the top cover right above that sensor. There may be a way to adjust that sensor alignment with the cover installed for final checkout. If could find some old instructions MIGHT try realigning it.
The Gesswein is a true emulator- it was designed to be able to image a failing drive then remove that drive and emulate it with the image you read, the docus claim to support emulation on a ?Rainbow! You should test if the Rainbow can boot replacing the HD with the Gesswein emulating the HD. If it won't boot, or only boots MS-DOS/CPM you would know the image isn't 100% good.
Just want to say I LOVE watching your videos. You are so engaging and brilliant at explaining these machines, and it's a joy to ride along on the troubleshooting journeys. I really appreciate the absolutely huge amount of work you put into making it seem so easy.
Interestingly, some of the strings in the Rainbow's drive for the startup menu seem to have ANSI escape sequences in them ( "[;H" moves the cursor to row 'n' colum 'm') which must be used to lay out the menu on the screen. Some of them also look to be dollar terminated (ending with a '$') meaning they are probably intended to be passed to standard output in MS-DOS, and are then processed by ANSI.SYS (or some variant thereof). However, that would require DOS to be loaded in the first place, so I wonder if this functionality is supported directly in the BIOS...
I didn't know that DOS also used $-terminated strings. But I know CP/M does. So maybe DEC just used what was common back then. CP/M 3 introduced a BDOS call to change the terminator, sou you could use 0-terminated strings instead.
Quantum drive with that 'wind gauge' is kinda neat. I bet that black plastic arm is detecting the 'wind' near the platter surface and is an 'unlock' signal to let the heads move from the landing zone. I notice a 'void sticker' on the cover right above that optical sensor for your plastic piece. Bet there's some way to adjust/ align that sensor with the cover installed for final checkout. If you could find some alignment instructions and put a scope on some head output, you have a chance to realign it???
I removed one out of a DEC rainbow - a Type 1 MFM, 10 Mb of capacity, along with its controller card. I put it in a very early AT machine and formatted it - it worked, zero bad sectors. I was blown away that it actually worked, being from 1981, it's older than me.
I remember back in those days, around the middle 1980s, MFM drives were common in PCs, larger drives were ESDI or sometimes SCSI. Drives were very expensive so sometimes we would scrounge up broken drives and have them repaired. Broken drives would sell for around $1 per megabyte, crazy by today's prices. My first PC had only floppy drives, I later bought a 20MB hard drive kit for it for $450 from a tiny little store in Austin, TX called Compuadd. I eventually added a 2nd 20MB drive to that PC, then replaced both with a 60MB RLL drive. My next PC was a 386 with 120MB IDE drive. Then I built another machine and eventually it had dual 383MB ESDI drives. That was in the early 1990s. Today my desktop computer has a 1TB SSD and a 4TB SSD. The 4TB SSD is 200,000 times the size of my first MFM drive.
And today 24TB drives are readily available. One company makes a 30TB drive and has even developed platters in the lab that will allow for 50TB drives soon.
Use the TAB key to enter complicated stuff and navigate through directories much more easily. Type the first few characters and hit TAB. The console will fill in the command or directory name as far as it is unique. If the command or name isn't fully autofilled, hit TAB twice to see a list of the possibilities. Then enter the next character and hit TAB. For example cd /opt/mfm/ TAB TAB will show you all the subfolders in /opt/mfm so you can add them while typing the cd command. You don't need to ls any more! No typos, fast and it also works with parameters for more modern commands.
A small tip for old MFM disks usage - they must be parked before power off, otherwise heads lands on usable platter area and scratches it. For Seagate drives parking area is seek to few tracks above maximum usable format (most close to center of disk). For others disks maybe different, some later disks have separate command for parking operation accessible by controller BIOS.
In my opinion "spinning" rust was as influential in early computers as the transistor. Storage was always the biggest issue with computers in the 60's and 70's, and magnetic media was the answer. We are so spoiled today with super cheap mass storage.
Back in the early 90's (maybe around 1993 or so?), I had an old 5 1/4" full height 80 MB drive. Looong before UA-cam, and I hadn't even heard the word "internet" yet. I was very curious, so I popped the top off the drive and ran it for quite some time, bare. I don't remember much, just that it was built before voice-coil head movement, so it had a stepper motor to move the heads. The whole desk vibrated when the drive was seeking, and it fascinated me to watch it dance across the platters while booting, loading programs, etc. (However, my first hard drive was a full height 20 Meg drive on an XT. That could make the desk shake too!)
In linux, "cat filename| less" can be replaced with "less filename ", which is slightly better too. I was hoping you could read the image files as a filesystem, partly because I too have cpm floppies to read. You can make a block device from the file using"losetup" (though "mount" will do it for you when it knows it should), but i don't know of a cpm partition table or filesystem for Linux 😢
And if you desperately want to "cat" the rainbow file, you can always "lolcat" it. You may need to install lolcat first. This makes the rainbow file really rainbow!
I always use kpartx to create loop devices. This requires a valid partition table on the beginning of the disk though. But if it is a "superfloppy" aka no partitions, only filesystem then you don't need to create loop devices at all. Mount is able to handle it without
The linux program is cpmtools,,,, cpmcp, cpmls and so on.. You likely have to set CPMTOOLSFMT to the correct disk format, or nothing will happen (there is a cli switch as part of the command) But it is unlikely that it is a cp/m filesystem on that hd Tab completion save a lot of typing, also typing errors Most terminals have a copy and paste, but you can do it the hard way if you wish
Thanks Dave. I'm one of many people who will likely never see the USA, so your reports of those events are valuable to your many overseas supporters and viewers. 🙂👍
Back in the late 90s/early 00’s, a semi popular PC mod was to crack open your hard drive, cut a plexi window into the lid, and add some LEDs to watch the harddrive work. Pretty high risk though lol. You would turn off your AC/Heat, run a super hot shower to get the bathroom steamy, then wait for the steam to dissipate to make a pseudo clean room. Despite our best efforts, we still killed many a drive lol. That CDC Wren II is SCREAMING to have a plexiglass plate replace the lid lol
This was such a cool and interesting episode for me as I was weaned on ye olde MFM drives, back in the 80's-early 90's. At my first tech job I cobbled together an 8088 system, and was actually running a 5 megabyte Seagate for some months with the cover off(!). Never had a problem and only stopped using it when I upgraded.
I remember people putting a tiny piece of tape on the grounding strap. In fact they used to put a little square of plastic on that. I used to call it an inertial dampener. 😂
Once again, I live your emotions and facial expressions :D That's 'live' not 'love' because I do the self same things with my repairs. I've been meaning to buy a Gessweinmulator for years so I really should now. Cheers David.
The "bearing noise" on those two noisy drives may be the "antistatic tab" that rides on the end of the shaft. If so, a small drop of oil will shut it up. See 26:34 for an example of the antistatic "brush". It is the copper coloured metal shim that sits over the end of the shaft, and touches it to dissipate any charge build up from the spinning motor and the heads. In the past some people simply removed those or bent them out of the way, but I suggest a little oil or conductive grease on the contact point might be a better solution. EDIT: You figured it out. Was a very common problem back in the day. EDIT2: See if you can find the service manual for the RD52 I'm pretty sure there was a way to do a low level format on those, that re-created the servo tracks.
As far as I can recall, the Winchester disk drives did not have servo platters, the track and sector positions where encoded as preamble data when low level formatted. Also in those 5" drives there where both MFM (Modified Frequency Modulation) and RLL (Run Length Limited) types and they used different types of controller boards, they were not interchangeable.
Last opened drive: The plastic arm resembling more modern drive arms swinging out is a vane that detects air movement which can only be present if the platters are spinning. It works against a spring, and typically mechanically unlocks the heads, and in some cases moves an optical interruptor that tells the drive it is safe to have the heads out. The inertia I'm sure you're aware of the stories but they used to write code that would make the really large drives walk.
Oh, man...what an EPIC job you have done here! Thanks for the shout-outs, and the channeling (I hope it helped! Ha!) but more importantly, congratulations on tackling this entire project!
Drive spinning down shortly after spin up could be an RPM failure. If the drive doesn't reach it's expected RPM with a specified amount of time it's spun down by the drive's onboard controller as it's considered defective.
Yup, can be caused by either mechanical or electronic problems: Either the spindle bearings have too much friction to get the drive up to full speed or there's some power related issues. I had driver issues, where the spindle motor was missing a phase, or simply bad capacitors causing the spindle motor not to have enough power to reach the correct speed.
That squeaky scrub of the first drive’s spin up through a set of AirPods reached into my hindbrain and brought back many memories of the back room of the PC repair shop in the early 90s as these things were dying in the last few old 8088’s that came in. I remember having to use debug in DOS to hit the low level format on my Winchester controller. Now, I have 6tb of casual storage on my desktop with 80tb in the basement for cold storage. 10mb as a kid in the 80s…
I remember the ST-506/412 interface drives, and MFM, on my early PC. I had a ST-4096, which was a whopping 80 megabytes! It was full height when half-height had already become normal for PCs. It was released in 1986, but I got it a bit later. It was so heavy that it had a reputation for browning out power supplies when spinning up -- people would have to hit the reset button after they came to speed to boot. Now I'm feeling nostalgic for terminating resistors. Yesterday I was fighting USB cables and WiFi on an old laptop. Different monkeys, same song keeps repeating.
6:24 The ST-4097 used RLL encoding. The 4096 was MFM, and the 4097 (85 MB) was the same drive hardware updated to use RLL. But it is only 71MB? That's less than the 4096. And it's 17 sectors, not 26. So it seems to have been formatted for compatibility with MFM drives of similar specs. As for the loud spindle: I remember it was normal to do percussive maintenance on them. Just give the protruding stub of the axle a little tap. I remember a co-worker had a drive that started to sound like an airplane was trying to land on his PC. MS-DOS would start the director listing with "Volume in drive C: is " and he named to drive "Very Loud".
In the default terminal of most linux distributions, bash, you can press the up/down arrows to select a command that you previously ran and the enter key to run it. The up arrow selects the next one in the list of previous commands, the down arrow goes the other way in the list.
@@alexanderdelguidice4660 also ctrl-r will start searching the history. Press that then start typing a command and it will try and find the best match in your history to fill the rest of it.
If you're the kind of person the deals with old drives that sometimes need open inspection and servicing then I would suggest considering buying a clean box to do this type of work. You can open drives to see their internal condition without exposing the platters and internals to unfiltered air and potentially ruining the whole drive. Clean boxes cost maybe a couple of thousand dollars and are meant for exactly this type of work.
Lots of good memories here. I recall having to do maintenance on multiple drives with the ground strap cleaning, there was also a little thin oil applied to some of the bearings. My memory is a bit fuzzy here, but weren't there some termination resistors and jumpers on the drives when they were daisy chained onto a bus. Some of those drives may be set up wrong for a single drive config.
The oil on the base plate there is some of the oil present on the platters. Whenever you have a drive with "stiction" problems, it's because the heads have settled into that oil and become stuck.
That first drive looks like a CDC/MPI/Imprimis one, particularly a Wren II which was a very advanced drive for the time. They were made with ST506 (they called it CDC506), ESDI and ATA interfaces. They were apparently the first company to ship a drive with an ATA interface. You are fortunate to have two which don't have bad tantalum capacitors. If they have Seagate branding it means they were made after Seagate acquired Imprimis in 1989.
7:15 Do these drives automatically park their own heads on power loss? Asking because back in the day when I had a full height drive (i dont rememberwhat type it was) I had to run a park program before powering off.
Yes, there's the ST-251 and it's other brethren (ST-277R, ST-296N, ST-251N...)that automatically parked the heads. I think the ST-4096 also did (?) since it had a voice coil mechanism with a return thing
I recall using the PARK command as a precaution. The problem is that if your drive has just lost power, it has no power to park the heads. Drives that had voice-coil positioning would have a large capacitor to park the heads with its "last gasp". This feature was actually a problem on some large format drives, one 14" drive that I used to work on was able to do an emergency retract while the engineer was doing an alignment on it. This broke the alignment tool, misaligned the head you were adjusting and, if you were in the wrong place, broke your finger.
@@ub321 Yes, quite a few of them did. The ones with stepper motor actuators would use the inertia of the platters to generate electricity to park the heads. They would make a BRRtrtrtrtrtrtrtrtr sound as they spun down.
@@Alexis_du_60 The ST-251 used a stepper motor, I still have one and you can see the motor from the underside. Voicecoil actuated drives were very rare in that era, the ones in this video must have been quite high end units.
Were you the one to cut the warrantee stickers? If not, somebody could have been in there before... Still, thanks for the fun tour of various models of spinning rust!
22:57 that plastic piece not only has start/end markers but also the servo information on it (the grey-looking circle-segment) so without that in proper alignment, no tracks will ever be found by the drive again
The drives with status not coming good that don't spin down are likely ok, but can't communicate. The most likely IC to be the problem, because it's the exposed part (esd-wise, and surge wise), is the line driver that drives the index, etc pin on the edge connector. Back-track across the pcb with a continuity tester, that gets the right pin of the right chip, and that almost always on old through-hole boards will be a standard part. Look up chip pinout, use scope to test inboard signal. Simples. (It that's not it, without schematic, it's too much work.) Replace IC, working drive. And yeah those grounding buttons were carbon originally, and yes they ended up squealing by the early 90s. Clean and done. I think I did about 20 MFM drives for some friends back in 1993. They paid me with a drive.
the manuals for those old drives are awesome! they have all sorts of schematics and test procedures. i'm sure if you find one every one of those drives can be fixed. even the one that is probably out of alignment now.
WD_3B1 would be for the AT&T 3B1 UNIX computer. TANDY_16B is the Tandy Model 16 Xenix computer. A blast from the past having worked on both machines at both companies.
Hey, I love your content. As a suggestion and a tip: Pressing the up and down keys in a Linux shell will let you scroll through previously entered commands. It might save you a bit of typing repetitive and similar commands
I have been watching your channel for quite some time now and it is safe to say that who you are and what you do is abolutely frikkin' awesome. The only thing right now that saddens me is that I'm already oversubscribed with Patreon related stuff (and yeah, I have this problem whereby I can't see on the Patreon page who exactly I'm patreoning apart from maybe 4 and I know it is much more than that - and I know I will have to deal with this at some point - I think the youtube Join thing might work better). David, you are just amazing and I'm hoping some day that you'll get the tape reader working on your improved vacuum tube microcontroller. In my spare time I like doing retro coding - my preferred retro coding for a few years is stuff relating to the Atari 8 bit line of computers (preferably, Atari800Xl / 65XE). I like coding for that machine because in Jeff Minter's words it is the ultimate lazy programmer's computer. Eh, and it also is a mite faster than the c64, although the Antic chip does slave the CPU, just as the VICII does on the C64. One thing I did discover recently is that the MADS assembler will not create .prg files, so I guess I'm going to have to use the KICK assembler instead. I want to write a game for the Commodore PET to start out with and it will involve a woodchipper - which I call a banana machine. Yeah and also in Ollish - "Uma flominap untsanj" After I write a game for the PET, I'll write something for the VIC 20, then for the C64 and then the Plus4. Maybe at some point I'll try to get into Z80 coding. Oh - back in the day I did write a program on my ZX81 that involved just poking stuff into a REM statement that inversed the characters on the screen. Didn't have an assembler then for that, just made use of Rodney Zak's book. The only other thing is - if only the Sinclair RAM pack used SRAM instead of DRAM. DAMN. One last thing - has anyone managed to compile the Turbo Rascal Syntax Error on the raspberry Pi 5? It ain't a problem per se, I'll be happy just to assemble without any libs at all - I mean...I don't even use macros in my assembly code. Heh. My mode of speech varies, but I'm from the UK and right now, since 1993, I live in Northampton UK. Yeah, where Alan Moore lives. Sorry about the rambling prose here - I just wanted to let you know that you are amazing. One last thing - with respect to rabbits - have you ever seen the 1979 version of "Watership Down"? Back in the day I went to the cinema to see that.
Absolutely can recommend the Large Scale Systems museum. Went there 2y ago and it was awesome. Great people, Great museum, and I am really upset that I have yet to do my google review, nor ordered my tshirt.
The ground strap on that drive may or may not have even been needed. Many drives actually had an alternate ground path to avoid static buildup, and the "fix" for the horrible squealing drive was simply to rip off the strap! However, I'm not sure whether that applies to your drive.
My old Arche Rival came with a 40 MB MFM hard drive. She was a 386 unit. Had a power switch that could break fingers. Upgraded she ran DOS 6.22 and Windows 3.11b. Starting up the PC was an adventure into sound. Most of the "Whirrr" sound was from the drive. Current Arche Rival I found on eBay seems to have a current IDE drive. Will have to delve deeper into the PC to learn more. Will need a UPS and power conditioner. Scary tantalums all over the board.
I remember that sound from your disks. I had a 70MB disk with the same sound decades ago. On the bottom of the disk spindle, I think it has hidden under the PCB, there is a grounding leaf spring touching the spindle. That was making the sound. Try lifting the spring to hear if the sound goes away. If it does, try bending the spring a little so the spindle touches another part of the spring.
34:06 reminds me a bit twisting the dragon's tail, where the dudes played with radioactive cores, trying to get them going critical without killing themselves. they used a screw-driver to distance something. it did not work. well, not killing them did not work. going critical worked quite well.
It's a little bit dangerous to connect live power supplies to drives. The Molex plug has no mechanism for affecting which pins touch first (unlike many modern connectors which have longer ground pins). This means you can have voltage traveling though circuitry in unexpected ways until full contact is made, such as 12V flowing through devices to get to 5V in the absence of ground, or voltage flowing to connected electronics in ways that exceed their ratings.
I was expecting to see a storm of messages in the comments section about this. Now I see only your message and it's sad that common sense from the past is now forgotten. As a side note I did burn my Quantum 120MB drive, just by testing what will happen if I don't follow the directions :D
Working with old MFM drives is "interesting", I can verify. I replaced the drive in my 80s UNIX machine with a better MFM drive and it took literally days to get all the bad blocks marked and the drive formatted before I could even get to the installation of a large box of floppies. Nice video illustrating working with very old hardware with spinny bits
Haven't heard a word about MFM drives since I upgraded an IBM XT with one from one of the first big box computer stores in the 1990s. These early stores were a wealth of information and tips since this was pre-internet. If I'm remembering correctly the drive was a full height with 10MB of storage. This was back in the 640 KB RAM days.
In My day, when we had to open up a Modified FM Drive, and if it would not recondize the "index" tracks, we would low level format them. both RLL and MFM controllers had a low level format utility on the ROM it self which you could invoke by (if on Dos) using the GO command on Debug "GO 1111:ffff" kinda situation. please check that.
I didn't realize that there were some of these really old personal computer drives that already had a voice coil instead of a stepper! Also interesting to see a phase with stepper/swinger combo instead of voice coil.
And to think, I had a basement FULL of original working IBM AT/PC/XT machines, with monitors and keyboards. All of it eventually got recycled. The collection would have probably been worth a lot of money today.
I remember as a teenager these MFM drives ware a dime a dozen, as in they ware free. I had a lot of fun with them. About the RD52-A the black plastic arm that moves when the drive starts and stops, that the drive head lock. As the drives spin-up the air pushes on that arm moving it away to to unlock the drive heads and obviously the reverse as the drives spin down. This made the old "Park head" commands unnecessary.
Great episode! Since you successfully backed up the data from two of the drives I would have loved to see you lay the Rainbow bits back down on a different drive and bring the Rainbow up on different spinning rust (or even boot using the MFM Emulator if that's an option?). I've never used the MFM Emulator, but the errors you saw at the end of the Rainbow backup remind me of errors that the Greaseweazle displays when it encounters certain types of copy protection. I don't believe the copy protection scheme is transferred in those cases so actually testing your Rainbow backup might be a good idea. Thanks for all your awesome content!!!!
I have a Rainbow and did just this. I used the MFM emulator to image the disk and now use that emulated disk in the machine. It works just the same. The MFM emulator is fantastic. I need to get a 2nd so I can image my Pro-350 disk.
I wonder how well that power supply you were using was doing? Some supplies need a normal current draw on the +5V to throttle up the switcher to provide rated current on the other voltages (like the +12V). Some drives need quite a bit of +12 and that Beaglebone probably isn't drawing much on the +5V.
Interesting info: For this era of drives, the first digit in the drive model number is often the number of platters or heads, eg: the ST-412 has 4 heads.
This is not correct. ST-506 also had 4 heads. ST-251 got 6 heads. Actually, Seagate used numbering system where first digit means form-factor, i.e. 4 = 5.25 full height, 2 = 5.25 half height, 1 = 3.5 half height, 3 = 3.5 slim (1 inch height)
The second drive you opened looked like it had damage on the top side of the second platter from the top of the stack. At 26:37 the large chip on the left is an 8039 microcontroller made by National Semiconductor. The chip on the right is an MB673304U made by Fujitsu. I couldn't find a datasheet for it. I don't know what is on your Linux cheat sheet. The lsblk command is a quick way to see the drives seen by the system. I use df to check what partitions are mounted. Instead of doing a cat of a file and piping it to less you can ask less to open the file directly (less rainbow.txt). If you didn't want to make a file with the strings from a file you could just run "strings filename.txt |less". The command man will give you information about using commands. To help find commands you need you can use man -k, or apropos.
Thanks for all the great videos! I'm not sure but I don't think that plastic thing goes over the metal part on that drive. I think it sticks out and goes into that black plastic sensor area. It's got lines on it like an optical encoder. Like on an old ball mouse. I don't know how you'd align it properly but I suspect that's how the controller knows how much it's moved the voice coil or whatever. I might be wrong, as I've never seen this on anything but printers and ball mice. Mouses. Meece? Anyway... I can't know for sure but I bet that's what's going on. Just watched that segment again and you basically say all this so... okay! But thanks for the all great videos again!
On the drive that has the black piece that moves after the drive starts spinning up is a head lock. It locks the heads in place for transportation when the drive is powered down. The indexing tag was the only problem. Previous attempts at trying to fix the drive may have wiped the data or someone used a magnetic erase device on it and there is no reference data. The error could be from a setting on the drive itself that is making it act as if it is part of a two drive system and a jumper or jumper block will need to be moved to set it as drive 0.
I have fried a few different things by accidentally plugging in connectors off by one pin or row. It's extremely frustrating, usually happens when I'm tired.
Most of my mfm drives that use servo tracks no longer work. They behave similar to that cdc one. Those that just use stepper moters are still generally work.
WOMBAT. Webster's Omnipotent Mass-Storage Builder and Tester. I worked for Webster when that product was developed. The wide row of 8 ceramic fuse-link ROMs with the paper labels contain the 60 bit wide microcode. As well as all the hardware control, data transfer and QBUS control code, it also contains a 6800-inspired emulator. The EPROM contains WOMBAT, written in 6800 assembly. The full theory of operation and the schematic is in the User Manual, which you can find on-line if you don't have it already. The Microcode was (of course) assembled by Macro-11. The 6800 code was probably assembled the same way, except I remember a PC-based assembler being used later as it was a lot faster.
The first computers my former employer received were used DEC Rainbows with dual floppy drives in the early 80s. They were terrible. The available 10MB hard drives sold for $3k. We couldn't afford that since we got 30 PCs. 🙂
Usagi: "I am bad at UNIX"
Also Usagi: "Here is my PDP-11" :)
Totally going to install RSX-11M on it so I can be bad at multiple operating systems!
@@UsagiElectric Are you bad at RSTS already?
@@UsagiElectric A PDP-11 without a proper 2BSD on it is a Sad Thing(TM) :)
2.11BSD can be installed on PDP-11.
@@jurajlutter yep, I ment the entire 2BSD family, so 2.8 or 2.10 or even earlier would be an option, but obviously, I'd also run 2.11BSD (which I do on my PiDP-11 :)
It is true! We programmers make better lovers. In fact I made one just the other day, and it only crashed three times.
It's soo easy.
PEEN_LEN = 6;
REPEAT
IN PEEN_LEN;
OUT PEEN_LEN;
UNTIL CLIMAX;
Having it go down on you only three times must have been disappointing
At least your stack still overflows
This works for me:
alias sex "updatedb; locate; talk; date; cd; strip; look; touch; finger; unzip; uptime; gawk; head; apt-get install condom; mount; fsck; gasp; more; yes; yes; yes; more; umount; apt-get remove --purge condom; make clean; sleep"
Indeed. the proof is the film Weird Science
The build quality is crazy. Its so robust, and bulky. Seeing old technology is so cool.
some programmer has been waiting decades for you to find that message!
some where some old guy got mystery wood when that egg dropped....it would be evdence of a just god if he got know it's been found and seen by the whole HELLORD world
Worth it.
No doubt 🤣
Hey there... don't give up on the drive with the alignment optic thing quite so fast.... I don't think those drives have any permanent index data on them, if everything is working, you might well find the a simple low level format will fix it... I did this once back in about 1994 for a bet, and actually removed the heads and all the platters from a drive, literally cleaned them with window cleaner spray, put the thing back together and got it fully working again after a low level format... where, in fact, no bad sectors were found,
That reminds me how I revived a Miniscribe 10 MB hard drive which had internal contamination. I cleaned the platter with Windex or something like that, I gently cleaned the heads with IPA, did a low-level format and the drive was good to go with just a few KB of bad sectors. That's a good thing that old MFM drives are so stupid, as if they're not too much damaged they can be brought back to life with relative ease and without special tools.
Fun fact: the platter was not coated with the iron oxide typical of old 5.25 MFM drives, but was already an anodized one. That's quite odd for a drive so old.
I kind of remember the low level format being in the ROM of the drive and there was some way you just ran it. That's going back to the first drive I purchased in 1987 so might be a little off there.
@@JimHarrigan IIRR at c800:5 but who knows with this odd adapter card
Probably, but then you just erased whatever was on it, which might be interesting. I'd intercept the alignment sensor to see what it sees and then manipulate the signal to get back into alignment. Oscilloscope and a poti should be enough for that.
@@JimHarrigan On later IDE drives there was critical structural data, firmware for the drives microcontrollers etc stored on the drive itself, so you could not low level format it... but older drives were very basic.
28:55 absolute gut laugh when the drive spun down right in the middle of your victory lap. Sorry, I'm a bad person. Thank you for leaving it in 🤣🤣
You know, you voided your warranties on those drives you opened.
I think they may be out of warranty by a little bit.
@@RichardFraser-y9t Just a little tiny bit - not even a century
I actually took footage of me slicing the warranty sticker on all three drives, but it didn't make the final cut.
@@UsagiElectricI don’t know why, but whenever I need to open up old electronics, those “warranty void” stickers always make me hesitate, even though the warranties expired decades ago.🤪
@UsagiElectric with the exception of things like phones (where you can't do anything once you're in), I always slice the stickers. even if I'm not opening the thing up, stickers are destroy-on-sight.
About 30 years ago when I was a teenager I took the cover off of a working 20MB MFM drive in an XT (they were worthless old junk at the time) and powered it up so that I could watch the heads move. It was fascinating but the drive started developing a bunch of bad sectors after only a few minutes of operating that way. I think that running a drive with the cover off disturbs the airflow and causes the heads to crash even if there is not a bunch of dust in the air so I would never power up an open drive I had any intention of trying to fix. I remember disassembling several other 10-20MB drives back then because they weren't worth anything, really wish I'd saved all that stuff but that's how it goes. Old junk becomes desirable collectibles but you never know what will be desireable.
The problem is microscopic dust -- what always floats around that you can't see is nonetheless bigger than the gap between the platter and the floating head. Some dust lands on the platter and sticks, and the head hits it and bounces -- microscopically, but still enough to screw up the read and eventually damage it the surface, and the head. That's why data recovery is done in a clean room (if they need to open up the drive), with literally NO dust, microscopic or otherwise.
I once scrapped some ancient nonworking RLL drive, only to shortly learn that in the recovery market, the durn thing was worth (are you sitting down??) .... $900 !!! That's Nine Hundred Dollars, not a typo. ARGH!
I did the same! It starting making a lot of noise when a head hit a platter.
"I'm in a purely Linux invironment and I have no idea what I'm doing."
You did it! You broke Linux down to it's bare essentials!
Even knowing those drives were probably cooked, watching them be opened was difficult. 😅 It took me YEARS to learn to be OK with opening drives, even when known to be dead. But then I found the magnets, and all was good. Also, as a fellow auto enthusiast, I have far too much sympathy for the 10 minute job that takes 2 days. 👍
I opened up drives that misbehaved in the 90ies just to punish them. Also have some very strong magnets from "modern" drives (maybe 10 years old)
or even in the 80ies, not sure when the first drive made me mad
And some of the platters made beautiful wind chimes.
But yes, it was painful watching them opened, and knowing that now the platters and heads will have dust damage. They _might_ have large enough sector tolerances to still read, but for how long?
@@Reziac I wish I lived in an area I could turn some of my superfluous platters into wind chimes. Unfortunately, in the Appalachian region of Virginia, we either have no wind, or tropical storm force. (Already lost 2 wind chimes to the latter.) Still thinking about different crafty ideas for them. I like the iridescent coatings some of the platters use. Seems like a light sculpture opportunity of some kind.
That "bearing" noise actually sounds like noise the static eliminator makes on these old drives. The static eliminator strip is located outside the sealed compartment on the bottom side of the drive. Some times it's dirt and wear that makes them squeal. You can try using a little IPA on a piece of notebook paper dragged between the spindle and static eliminator. If not dirty, try slightly increasing the tension on the static eliminator. Well, I wrote this before watching the entire video. Glad you found it. I've heard that noise way too many times in days gone past.😅
Advice: You can look through your command history in Linux console with Up and Down arrows.
To add to that: you can use the tab key to auto-complete file paths.
Control r to reverse search through command history
and copy/paste in a terminal exists. :)
And type the command 'history' to see a full list of remembered commands. If you type your password into an open prompt by accident, you can use the same command to clear individual entries.
Thanks! Though I did already know that shortcut. When I'm doing screen capture for a video I always retype the full command so that it gives the viewer, especially those not familiar with Linux, time to visually locate the command and mentally parse what is being typed before it flies off the screen.
The glass plate in the RD52 is the servo position encoder.. If you look closely at that arc pattern you'll see tiny metallic stripes.
It's called the "Light Positioning System". These drives have a servo surface but it is actually possible to configure them to ignore the servo surface and just use the LPS for tracking.
I noticed a 'void warranty' sticker on the top cover right above that sensor. There may be a way to adjust that sensor alignment with the cover installed for final checkout. If could find some old instructions MIGHT try realigning it.
@@douro20 If you can track LPS, should you not be able to low level format the drive and have it function again?
@@mfree80286 Yes, as long as everything else in the drive is in good condition.
I had seen a similar mechanism in 3.5 SCSI drives (Quantum prodrive LPS), and didn't realize it goes back this far. pretty neat.
The Gesswein is a true emulator- it was designed to be able to image a failing drive then remove that drive and emulate it with the image you read, the docus claim to support emulation on a ?Rainbow! You should test if the Rainbow can boot replacing the HD with the Gesswein emulating the HD. If it won't boot, or only boots MS-DOS/CPM you would know the image isn't 100% good.
I was waiting for that and it didn’t happen. Why didn’t it happen? 😢
Exactly what I was thinking. The only way you can confirm the veracity of a backup is to write it to another device and see if it runs correctly.
Just want to say I LOVE watching your videos. You are so engaging and brilliant at explaining these machines, and it's a joy to ride along on the troubleshooting journeys. I really appreciate the absolutely huge amount of work you put into making it seem so easy.
Interestingly, some of the strings in the Rainbow's drive for the startup menu seem to have ANSI escape sequences in them ( "[;H" moves the cursor to row 'n' colum 'm') which must be used to lay out the menu on the screen. Some of them also look to be dollar terminated (ending with a '$') meaning they are probably intended to be passed to standard output in MS-DOS, and are then processed by ANSI.SYS (or some variant thereof). However, that would require DOS to be loaded in the first place, so I wonder if this functionality is supported directly in the BIOS...
I didn't know that DOS also used $-terminated strings. But I know CP/M does. So maybe DEC just used what was common back then.
CP/M 3 introduced a BDOS call to change the terminator, sou you could use 0-terminated strings instead.
Quantum drive with that 'wind gauge' is kinda neat. I bet that black plastic arm is detecting the 'wind' near the platter surface and is an 'unlock' signal to let the heads move from the landing zone.
I notice a 'void sticker' on the cover right above that optical sensor for your plastic piece. Bet there's some way to adjust/ align that sensor with the cover installed for final checkout. If you could find some alignment instructions and put a scope on some head output, you have a chance to realign it???
I'm just amazed by this drive seemingly using an optical alignment solution that involves a bit of plastic and ink :D
I removed one out of a DEC rainbow - a Type 1 MFM, 10 Mb of capacity, along with its controller card. I put it in a very early AT machine and formatted it - it worked, zero bad sectors. I was blown away that it actually worked, being from 1981, it's older than me.
I remember back in those days, around the middle 1980s, MFM drives were common in PCs, larger drives were ESDI or sometimes SCSI. Drives were very expensive so sometimes we would scrounge up broken drives and have them repaired. Broken drives would sell for around $1 per megabyte, crazy by today's prices. My first PC had only floppy drives, I later bought a 20MB hard drive kit for it for $450 from a tiny little store in Austin, TX called Compuadd. I eventually added a 2nd 20MB drive to that PC, then replaced both with a 60MB RLL drive. My next PC was a 386 with 120MB IDE drive. Then I built another machine and eventually it had dual 383MB ESDI drives. That was in the early 1990s. Today my desktop computer has a 1TB SSD and a 4TB SSD. The 4TB SSD is 200,000 times the size of my first MFM drive.
And today 24TB drives are readily available. One company makes a 30TB drive and has even developed platters in the lab that will allow for 50TB drives soon.
i know. i paid about $500 for a 20mb drive, and was happy to get it.
Use the TAB key to enter complicated stuff and navigate through directories much more easily.
Type the first few characters and hit TAB. The console will fill in the command or directory name as far as it is unique. If the command or name isn't fully autofilled, hit TAB twice to see a list of the possibilities. Then enter the next character and hit TAB.
For example
cd /opt/mfm/ TAB TAB
will show you all the subfolders in /opt/mfm so you can add them while typing the cd command. You don't need to ls any more!
No typos, fast and it also works with parameters for more modern commands.
I think that grounding strap is intended to have a carbon button on it and not metal-to-metal. That's where the black gunk came from.
A small tip for old MFM disks usage - they must be parked before power off, otherwise heads lands on usable platter area and scratches it.
For Seagate drives parking area is seek to few tracks above maximum usable format (most close to center of disk). For others disks maybe different, some later disks have separate command for parking operation accessible by controller BIOS.
Made me cringe when he just unplugged it.
In my opinion "spinning" rust was as influential in early computers as the transistor. Storage was always the biggest issue with computers in the 60's and 70's, and magnetic media was the answer. We are so spoiled today with super cheap mass storage.
Back in the early 90's (maybe around 1993 or so?), I had an old 5 1/4" full height 80 MB drive. Looong before UA-cam, and I hadn't even heard the word "internet" yet. I was very curious, so I popped the top off the drive and ran it for quite some time, bare. I don't remember much, just that it was built before voice-coil head movement, so it had a stepper motor to move the heads. The whole desk vibrated when the drive was seeking, and it fascinated me to watch it dance across the platters while booting, loading programs, etc. (However, my first hard drive was a full height 20 Meg drive on an XT. That could make the desk shake too!)
In linux, "cat filename| less" can be replaced with "less filename ", which is slightly better too.
I was hoping you could read the image files as a filesystem, partly because I too have cpm floppies to read. You can make a block device from the file using"losetup" (though "mount" will do it for you when it knows it should), but i don't know of a cpm partition table or filesystem for Linux 😢
And if you desperately want to "cat" the rainbow file, you can always "lolcat" it. You may need to install lolcat first.
This makes the rainbow file really rainbow!
Or better without the temp file: "strings filename | less"
cpmfuse supports CP/M filesystems using fuse (userspace filesystem). Or you can use cpmtools
I always use kpartx to create loop devices. This requires a valid partition table on the beginning of the disk though.
But if it is a "superfloppy" aka no partitions, only filesystem then you don't need to create loop devices at all. Mount is able to handle it without
The linux program is cpmtools,,,, cpmcp, cpmls and so on.. You likely have to set CPMTOOLSFMT
to the correct disk format, or nothing will happen (there is a cli switch as part of the command)
But it is unlikely that it is a cp/m filesystem on that hd
Tab completion save a lot of typing, also typing errors
Most terminals have a copy and paste, but you can do it the hard way if you wish
Thanks Dave. I'm one of many people who will likely never see the USA, so your reports of those events are valuable to your many overseas supporters and viewers. 🙂👍
Baby Bun is adorable.
Back in the late 90s/early 00’s, a semi popular PC mod was to crack open your hard drive, cut a plexi window into the lid, and add some LEDs to watch the harddrive work. Pretty high risk though lol. You would turn off your AC/Heat, run a super hot shower to get the bathroom steamy, then wait for the steam to dissipate to make a pseudo clean room. Despite our best efforts, we still killed many a drive lol.
That CDC Wren II is SCREAMING to have a plexiglass plate replace the lid lol
This was such a cool and interesting episode for me as I was weaned on ye olde MFM drives, back in the 80's-early 90's. At my first tech job I cobbled together an 8088 system, and was actually running a 5 megabyte Seagate for some months with the cover off(!). Never had a problem and only stopped using it when I upgraded.
I remember people putting a tiny piece of tape on the grounding strap. In fact they used to put a little square of plastic on that. I used to call it an inertial dampener. 😂
Great job getting the data off those drives. Look forward to seeing you fix the others.
Once again, I live your emotions and facial expressions :D That's 'live' not 'love' because I do the self same things with my repairs. I've been meaning to buy a Gessweinmulator for years so I really should now. Cheers David.
The "bearing noise" on those two noisy drives may be the "antistatic tab" that rides on the end of the shaft. If so, a small drop of oil will shut it up. See 26:34 for an example of the antistatic "brush". It is the copper coloured metal shim that sits over the end of the shaft, and touches it to dissipate any charge build up from the spinning motor and the heads. In the past some people simply removed those or bent them out of the way, but I suggest a little oil or conductive grease on the contact point might be a better solution.
EDIT: You figured it out. Was a very common problem back in the day.
EDIT2: See if you can find the service manual for the RD52 I'm pretty sure there was a way to do a low level format on those, that re-created the servo tracks.
As far as I can recall, the Winchester disk drives did not have servo platters, the track and sector positions where encoded as preamble data when low level formatted. Also in those 5" drives there where both MFM (Modified Frequency Modulation) and RLL (Run Length Limited) types and they used different types of controller boards, they were not interchangeable.
Last opened drive:
The plastic arm resembling more modern drive arms swinging out is a vane that detects air movement which can only be present if the platters are spinning. It works against a spring, and typically mechanically unlocks the heads, and in some cases moves an optical interruptor that tells the drive it is safe to have the heads out.
The inertia I'm sure you're aware of the stories but they used to write code that would make the really large drives walk.
Oh, man...what an EPIC job you have done here! Thanks for the shout-outs, and the channeling (I hope it helped! Ha!) but more importantly, congratulations on tackling this entire project!
Drive spinning down shortly after spin up could be an RPM failure. If the drive doesn't reach it's expected RPM with a specified amount of time it's spun down by the drive's onboard controller as it's considered defective.
Yup, can be caused by either mechanical or electronic problems: Either the spindle bearings have too much friction to get the drive up to full speed or there's some power related issues. I had driver issues, where the spindle motor was missing a phase, or simply bad capacitors causing the spindle motor not to have enough power to reach the correct speed.
Another useful command is "od", basically octal/hex/etc viewer on command line. Pretty handy.
od -x being particularly helpful for getting a hexdump
"hexdump -C" (or "hexdump -Cv" to see all the zeros) is much better; od is from the stone age of Unix
Tres cool episode, Mr. UE. ありがとうございます
Thanks!
That squeaky scrub of the first drive’s spin up through a set of AirPods reached into my hindbrain and brought back many memories of the back room of the PC repair shop in the early 90s as these things were dying in the last few old 8088’s that came in. I remember having to use debug in DOS to hit the low level format on my Winchester controller. Now, I have 6tb of casual storage on my desktop with 80tb in the basement for cold storage. 10mb as a kid in the 80s…
I remember the ST-506/412 interface drives, and MFM, on my early PC. I had a ST-4096, which was a whopping 80 megabytes!
It was full height when half-height had already become normal for PCs. It was released in 1986, but I got it a bit later. It was so heavy that it had a reputation for browning out power supplies when spinning up -- people would have to hit the reset button after they came to speed to boot. Now I'm feeling nostalgic for terminating resistors.
Yesterday I was fighting USB cables and WiFi on an old laptop. Different monkeys, same song keeps repeating.
Love this! Remember my first teardown of some old SCSI drives like this. They were huge, weighed like 10 pounds each.
love it, the heights and depths and then heights again - of madness and spinning rust :)
6:24 The ST-4097 used RLL encoding. The 4096 was MFM, and the 4097 (85 MB) was the same drive hardware updated to use RLL.
But it is only 71MB? That's less than the 4096. And it's 17 sectors, not 26. So it seems to have been formatted for compatibility with MFM drives of similar specs.
As for the loud spindle: I remember it was normal to do percussive maintenance on them. Just give the protruding stub of the axle a little tap.
I remember a co-worker had a drive that started to sound like an airplane was trying to land on his PC.
MS-DOS would start the director listing with "Volume in drive C: is " and he named to drive "Very Loud".
In the default terminal of most linux distributions, bash, you can press the up/down arrows to select a command that you previously ran and the enter key to run it. The up arrow selects the next one in the list of previous commands, the down arrow goes the other way in the list.
@@alexanderdelguidice4660 also ctrl-r will start searching the history. Press that then start typing a command and it will try and find the best match in your history to fill the rest of it.
DOS prompt too, at least on Windows.
If you're the kind of person the deals with old drives that sometimes need open inspection and servicing then I would suggest considering buying a clean box to do this type of work. You can open drives to see their internal condition without exposing the platters and internals to unfiltered air and potentially ruining the whole drive.
Clean boxes cost maybe a couple of thousand dollars and are meant for exactly this type of work.
Lots of good memories here. I recall having to do maintenance on multiple drives with the ground strap cleaning, there was also a little thin oil applied to some of the bearings.
My memory is a bit fuzzy here, but weren't there some termination resistors and jumpers on the drives when they were daisy chained onto a bus. Some of those drives may be set up wrong for a single drive config.
@24:30 As an Eagles fan, I approve 🤣
Sports aside, I absolutely adore your content! I never miss a video
Cars, vintage tech, tinkering 💖 Thank you! :)
Im so happy you rescued the bun bun,
Great episode. Nice doorstops and boat anchors!
I was looking at them thinking, they'd make a great heavy base for a retro desk lamp.
The oil on the base plate there is some of the oil present on the platters. Whenever you have a drive with "stiction" problems, it's because the heads have settled into that oil and become stuck.
You did disengage the transport lock on the CDC drive right? :)
That first drive looks like a CDC/MPI/Imprimis one, particularly a Wren II which was a very advanced drive for the time. They were made with ST506 (they called it CDC506), ESDI and ATA interfaces. They were apparently the first company to ship a drive with an ATA interface. You are fortunate to have two which don't have bad tantalum capacitors. If they have Seagate branding it means they were made after Seagate acquired Imprimis in 1989.
7:15 Do these drives automatically park their own heads on power loss? Asking because back in the day when I had a full height drive (i dont rememberwhat type it was) I had to run a park program before powering off.
I was wondering the same thing. Were any MFM drives self-parking?
Yes, there's the ST-251 and it's other brethren (ST-277R, ST-296N, ST-251N...)that automatically parked the heads. I think the ST-4096 also did (?) since it had a voice coil mechanism with a return thing
I recall using the PARK command as a precaution. The problem is that if your drive has just lost power, it has no power to park the heads. Drives that had voice-coil positioning would have a large capacitor to park the heads with its "last gasp".
This feature was actually a problem on some large format drives, one 14" drive that I used to work on was able to do an emergency retract while the engineer was doing an alignment on it. This broke the alignment tool, misaligned the head you were adjusting and, if you were in the wrong place, broke your finger.
@@ub321 Yes, quite a few of them did. The ones with stepper motor actuators would use the inertia of the platters to generate electricity to park the heads. They would make a BRRtrtrtrtrtrtrtrtr sound as they spun down.
@@Alexis_du_60 The ST-251 used a stepper motor, I still have one and you can see the motor from the underside. Voicecoil actuated drives were very rare in that era, the ones in this video must have been quite high end units.
Were you the one to cut the warrantee stickers? If not, somebody could have been in there before...
Still, thanks for the fun tour of various models of spinning rust!
22:57 that plastic piece not only has start/end markers but also the servo information on it (the grey-looking circle-segment) so without that in proper alignment, no tracks will ever be found by the drive again
The drives with status not coming good that don't spin down are likely ok, but can't communicate. The most likely IC to be the problem, because it's the exposed part (esd-wise, and surge wise), is the line driver that drives the index, etc pin on the edge connector. Back-track across the pcb with a continuity tester, that gets the right pin of the right chip, and that almost always on old through-hole boards will be a standard part. Look up chip pinout, use scope to test inboard signal. Simples. (It that's not it, without schematic, it's too much work.) Replace IC, working drive. And yeah those grounding buttons were carbon originally, and yes they ended up squealing by the early 90s. Clean and done. I think I did about 20 MFM drives for some friends back in 1993. They paid me with a drive.
the manuals for those old drives are awesome! they have all sorts of schematics and test procedures. i'm sure if you find one every one of those drives can be fixed. even the one that is probably out of alignment now.
WD_3B1 would be for the AT&T 3B1 UNIX computer. TANDY_16B is the Tandy Model 16 Xenix computer. A blast from the past having worked on both machines at both companies.
Hey, I love your content. As a suggestion and a tip: Pressing the up and down keys in a Linux shell will let you scroll through previously entered commands. It might save you a bit of typing repetitive and similar commands
We had to clean the spindle ground on HP 792X drives back in the late 70s. Weird sounds were normal when dirty. Basicly dirt on a beating point.
Hard drives are absolutely marvelous pieces of technology.
I saw "MFM" and was suddenly in a nostalgic bliss. :D
I have been watching your channel for quite some time now and it is safe to say that who you are and what you do is abolutely frikkin' awesome. The only thing right now that saddens me is that I'm already oversubscribed with Patreon related stuff (and yeah, I have this problem whereby I can't see on the Patreon page who exactly I'm patreoning apart from maybe 4 and I know it is much more than that - and I know I will have to deal with this at some point - I think the youtube Join thing might work better). David, you are just amazing and I'm hoping some day that you'll get the tape reader working on your improved vacuum tube microcontroller.
In my spare time I like doing retro coding - my preferred retro coding for a few years is stuff relating to the Atari 8 bit line of computers (preferably, Atari800Xl / 65XE). I like coding for that machine because in Jeff Minter's words it is the ultimate lazy programmer's computer. Eh, and it also is a mite faster than the c64, although the Antic chip does slave the CPU, just as the VICII does on the C64.
One thing I did discover recently is that the MADS assembler will not create .prg files, so I guess I'm going to have to use the KICK assembler instead. I want to write a game for the Commodore PET to start out with and it will involve a woodchipper - which I call a banana machine. Yeah and also in Ollish - "Uma flominap untsanj" After I write a game for the PET, I'll write something for the VIC 20, then for the C64 and then the Plus4.
Maybe at some point I'll try to get into Z80 coding. Oh - back in the day I did write a program on my ZX81 that involved just poking stuff into a REM statement that inversed the characters on the screen. Didn't have an assembler then for that, just made use of Rodney Zak's book. The only other thing is - if only the Sinclair RAM pack used SRAM instead of DRAM. DAMN.
One last thing - has anyone managed to compile the Turbo Rascal Syntax Error on the raspberry Pi 5? It ain't a problem per se, I'll be happy just to assemble without any libs at all - I mean...I don't even use macros in my assembly code. Heh.
My mode of speech varies, but I'm from the UK and right now, since 1993, I live in Northampton UK. Yeah, where Alan Moore lives.
Sorry about the rambling prose here - I just wanted to let you know that you are amazing.
One last thing - with respect to rabbits - have you ever seen the 1979 version of "Watership Down"? Back in the day I went to the cinema to see that.
Absolutely can recommend the Large Scale Systems museum. Went there 2y ago and it was awesome. Great people, Great museum, and I am really upset that I have yet to do my google review, nor ordered my tshirt.
The ground strap on that drive may or may not have even been needed. Many drives actually had an alternate ground path to avoid static buildup, and the "fix" for the horrible squealing drive was simply to rip off the strap! However, I'm not sure whether that applies to your drive.
My old Arche Rival came with a 40 MB MFM hard drive. She was a 386 unit. Had a power switch that could break fingers. Upgraded she ran DOS 6.22 and Windows 3.11b. Starting up the PC was an adventure into sound. Most of the "Whirrr" sound was from the drive.
Current Arche Rival I found on eBay seems to have a current IDE drive. Will have to delve deeper into the PC to learn more. Will need a UPS and power conditioner. Scary tantalums all over the board.
I remember that sound from your disks. I had a 70MB disk with the same sound decades ago.
On the bottom of the disk spindle, I think it has hidden under the PCB, there is a grounding leaf spring touching the spindle. That was making the sound. Try lifting the spring to hear if the sound goes away. If it does, try bending the spring a little so the spindle touches another part of the spring.
That drive's commedic timing was on point
34:06 reminds me a bit twisting the dragon's tail, where the dudes played with radioactive cores, trying to get them going critical without killing themselves. they used a screw-driver to distance something. it did not work. well, not killing them did not work. going critical worked quite well.
It's a little bit dangerous to connect live power supplies to drives. The Molex plug has no mechanism for affecting which pins touch first (unlike many modern connectors which have longer ground pins). This means you can have voltage traveling though circuitry in unexpected ways until full contact is made, such as 12V flowing through devices to get to 5V in the absence of ground, or voltage flowing to connected electronics in ways that exceed their ratings.
I was expecting to see a storm of messages in the comments section about this. Now I see only your message and it's sad that common sense from the past is now forgotten. As a side note I did burn my Quantum 120MB drive, just by testing what will happen if I don't follow the directions :D
Working with old MFM drives is "interesting", I can verify. I replaced the drive in my 80s UNIX machine with a better MFM drive and it took literally days to get all the bad blocks marked and the drive formatted before I could even get to the installation of a large box of floppies. Nice video illustrating working with very old hardware with spinny bits
Haven't heard a word about MFM drives since I upgraded an IBM XT with one from one of the first big box computer stores in the 1990s.
These early stores were a wealth of information and tips since this was pre-internet.
If I'm remembering correctly the drive was a full height with 10MB of storage. This was back in the 640 KB RAM days.
That bunny is so so cute!!!
In My day, when we had to open up a Modified FM Drive, and if it would not recondize the "index" tracks, we would low level format them. both RLL and MFM controllers had a low level format utility on the ROM it self which you could invoke by (if on Dos) using the GO command on Debug "GO 1111:ffff" kinda situation. please check that.
I didn't realize that there were some of these really old personal computer drives that already had a voice coil instead of a stepper! Also interesting to see a phase with stepper/swinger combo instead of voice coil.
21:47 I'll always remember the sound of the Quantum Fireball drives. I still have a few that work.
And to think, I had a basement FULL of original working IBM AT/PC/XT machines, with monitors and keyboards. All of it eventually got recycled. The collection would have probably been worth a lot of money today.
I remember as a teenager these MFM drives ware a dime a dozen, as in they ware free. I had a lot of fun with them. About the RD52-A the black plastic arm that moves when the drive starts and stops, that the drive head lock. As the drives spin-up the air pushes on that arm moving it away to to unlock the drive heads and obviously the reverse as the drives spin down. This made the old "Park head" commands unnecessary.
tektronix 6130 was a unix box tek made in their cambrian explosion era. would be interested to hear if anything interesting was on that disk.
Great episode! Since you successfully backed up the data from two of the drives I would have loved to see you lay the Rainbow bits back down on a different drive and bring the Rainbow up on different spinning rust (or even boot using the MFM Emulator if that's an option?). I've never used the MFM Emulator, but the errors you saw at the end of the Rainbow backup remind me of errors that the Greaseweazle displays when it encounters certain types of copy protection. I don't believe the copy protection scheme is transferred in those cases so actually testing your Rainbow backup might be a good idea. Thanks for all your awesome content!!!!
I have a Rainbow and did just this. I used the MFM emulator to image the disk and now use that emulated disk in the machine. It works just the same. The MFM emulator is fantastic. I need to get a 2nd so I can image my Pro-350 disk.
I wonder how well that power supply you were using was doing? Some supplies need a normal current draw on the +5V to throttle up the switcher to provide rated current on the other voltages (like the +12V). Some drives need quite a bit of +12 and that Beaglebone probably isn't drawing much on the +5V.
This was a really fun video. I always wondered what a hard disk drive looked like on the inside while operating.
16:53 Oh no! you voided your warranty! :D
Interesting info: For this era of drives, the first digit in the drive model number is often the number of platters or heads, eg: the ST-412 has 4 heads.
This is not correct. ST-506 also had 4 heads. ST-251 got 6 heads. Actually, Seagate used numbering system where first digit means form-factor, i.e. 4 = 5.25 full height, 2 = 5.25 half height, 1 = 3.5 half height, 3 = 3.5 slim (1 inch height)
The second drive you opened looked like it had damage on the top side of the second platter from the top of the stack. At 26:37 the large chip on the left is an 8039 microcontroller made by National Semiconductor. The chip on the right is an MB673304U made by Fujitsu. I couldn't find a datasheet for it.
I don't know what is on your Linux cheat sheet. The lsblk command is a quick way to see the drives seen by the system. I use df to check what partitions are mounted. Instead of doing a cat of a file and piping it to less you can ask less to open the file directly (less rainbow.txt). If you didn't want to make a file with the strings from a file you could just run "strings filename.txt |less". The command man will give you information about using commands. To help find commands you need you can use man -k, or apropos.
Thanks for all the great videos! I'm not sure but I don't think that plastic thing goes over the metal part on that drive. I think it sticks out and goes into that black plastic sensor area. It's got lines on it like an optical encoder. Like on an old ball mouse. I don't know how you'd align it properly but I suspect that's how the controller knows how much it's moved the voice coil or whatever. I might be wrong, as I've never seen this on anything but printers and ball mice. Mouses. Meece? Anyway... I can't know for sure but I bet that's what's going on. Just watched that segment again and you basically say all this so... okay! But thanks for the all great videos again!
On the drive that has the black piece that moves after the drive starts spinning up is a head lock. It locks the heads in place for transportation when the drive is powered down. The indexing tag was the only problem. Previous attempts at trying to fix the drive may have wiped the data or someone used a magnetic erase device on it and there is no reference data. The error could be from a setting on the drive itself that is making it act as if it is part of a two drive system and a jumper or jumper block will need to be moved to set it as drive 0.
I have fried a few different things by accidentally plugging in connectors off by one pin or row. It's extremely frustrating, usually happens when I'm tired.
That's so cool it emulates the mfm hard drive as an mmc storage device. I only do sd's and nvme's in linux! We're so spoiled now.....
Those were the days, when you could fit a burrito between platters on a hard drive.
14:19 you can save yourself time by just typing strings | less. No need to output it to a separate text file or use cat.
looks like the optical encoder on that drive with the grating that fell off it has an adjustment screw?
Most of my mfm drives that use servo tracks no longer work. They behave similar to that cdc one. Those that just use stepper moters are still generally work.
I recall an old drive that had a horrible screech. Turned out to be the cylinder grounding brushes and not the bearings
I have two MFM hard drives here. They might work, they might not, but as long as I don't test them, they might!
WOMBAT. Webster's Omnipotent Mass-Storage Builder and Tester. I worked for Webster when that product was developed. The wide row of 8 ceramic fuse-link ROMs with the paper labels contain the 60 bit wide microcode. As well as all the hardware control, data transfer and QBUS control code, it also contains a 6800-inspired emulator. The EPROM contains WOMBAT, written in 6800 assembly. The full theory of operation and the schematic is in the User Manual, which you can find on-line if you don't have it already. The Microcode was (of course) assembled by Macro-11. The 6800 code was probably assembled the same way, except I remember a PC-based assembler being used later as it was a lot faster.
The first computers my former employer received were used DEC Rainbows with dual floppy drives in the early 80s. They were terrible. The available 10MB hard drives sold for $3k. We couldn't afford that since we got 30 PCs. 🙂
St-506 is a 5mb drive. I have 4 of them.
that has to be the biggest BeagleBone cape I have ever seen.