Trying (and Failing) to Restore a 1970s CDC 10MB Hard Drive
ฝัง
- เผยแพร่เมื่อ 10 ก.ค. 2024
- VCF East (link below) is coming up soon, and my original plan was to take the Mini Centurion to VCF with me, but it’s just too slow. VCF is all about the best of the best, so I need to take a genuine Hawk drive with to demonstrate the Centurion in it’s proper glory. That means I need to restore another Hawk drive though, and that’s no easy feat. Conversely, it’s incredibly difficult and kicks my butt six ways to Sunday!
If you’re interested in VCF East, check this link:
www.eventbrite.com/e/vcf-east...
vcfed.org/events/vintage-comp...
Centurion Wiki:
github.com/Nakazoto/Centurion...
If you want to support the channel please hop over to Patreon:
/ usagielectric
Also, we now have some epic shirts for sale!
my-store-11554688.creator-spr...
Come join us on Discord and Twitter!
Discord: / discord
Twitter: / usagielectric
Intro Music adapted from:
Artist: The Runaway Five
Title: The Shinra Shuffle
ocremix.org/remix/OCR01847
Thanks for watching!
Chapters
0:00 Introduction
2:29 I feel the need, the need for speed
4:04 Installing the Hawk into the new 19-inch rack
6:17 Installing the Hawk into the new 19-inch rack again
8:48 Installing the Hawk… well, you get the idea
9:37 I have a platter problem
11:53 Repairing the Hawk
13:03 Cleaning up the heads and getting a new platter in place
14:55 Time to purge it, then spin it up
16:59 Victory! Almost at least
18:56 Trying to align the heads
19:59 Okay, that’s not going well
22:27 Bad heads, yup, plural, multiple heads… it’s one of those days
26:40 Bunny! - วิทยาศาสตร์และเทคโนโลยี
You still have a few failure points between the actual head coils and the connector. You should measure ohms *at the head* because all the bendy cable and crimped connectors between it and the pcb connector can fail, and thin cables can corrode.
YES! I was hoping someone thought of this!
very likely this is the issue
"This is the Lockpicking Computer Geek, and what I have for you today is setting a DIP switch with a wave rack... 1 is binding, 2 is a false gate, 3 is there... Okay, let's try again to see if it was not a Fluke"
I was gonna suggest that you make a cutout in the upper front angle bar on that rack to clear the Hawk drive's panel, but drilling two holes was way simpler! Haha, that was just brilliant :)
Got a bit worried seeing you lift the heavy thing all by yourself though.
Niiiiiice platter collection there :) and good job on refurbishing the heads - pity it didn't turn out to work, but still a work of love. It can indeed get frustrating! Well... catch you next time :)
Awesome comment!
Emulate the drive, and remove the heads from the physical drive but still have it spinning with a partial head assembly still in place so you still get the kerchunka. Then split the signal so that commands get sent to both the physical drive and the virtual drive, but data is only read back from the virtual drive.
Which defeats the purpose completly.
This makes me appreciate the service contract we had for our MicroVAX systems back in the day at GE. They were expensive as all getout, but the repair genie would visit and everything worked afterwards. Best of luck.
I've seen the miracles you and the other members of the Centurion Squad have pulled off to get a lump of dead parts turned into two completely functional Centurion Minicomputers. I have every faith in you that you will pull this off. Never say die!
Those heads look so happy to be cleaned.
You'll get there, Nakazoto! Remember: You're destined to take SkyNet down 😂
sudo shred-fuz /usr/bin/skynet; sudo rm -rf /etc/skynet
There is no fate but what we make ;-)
I bet the problem is the wires going to the heads are broken. Although coils can be burned by carrying too much current, it's more likely that the wires being under constant motion just broke the conductors. I would check the continuity directly at the coil-wire connection and see if it's still open circuit there. Wires are easier to replace than head transducers.
I was thinking the same thing!
@@Derundurel so was i, guess we'll know in the next episode.
Send it curious marc that mans channel is a wizard with crashed heads
Reminds me of the stuff Curious Marc does. His team tackles all kind of dinosaur computers. Not that he wouldn't touch more modern stuff on occasion. But this old stuff is intriguing. For someone growing up with C64 and following machines. (C128D, IBM PC, 486DX... Etc)
I'll give you credit. You definitely fight through. This is very challenging especially now that parts are hard to come by. Hope you are able to get this working. I'm sure you'll be celebrating if you do. We'll be celebrating right along with you.
You, sir, are doing yeoman's work on these old systems. Incredible patience that you have. Good to see younger people with such a penchant for learning and problem solving. Keep up the great work, my friend.
Hawk heads looks like a happy face.
I enjoy every single frame of your videos... I've no special knowledge on old computers (I started to follow you because the use of vacuum tubes...) but let me congrats with you for the effort profused in a search for solutions of the issues you have found in your Centurion project. Great work !!!!
I recall programming PDP 8 & 11 computers back in my student days, we sure didn’t dare take them apart back then! It’s fascinating watching you fearlessly restoring these old beauties, I’m cheering you on in the background! It’s surprising so many heads failing in the same drive, are they prone to failure or is there another cause?
You just did a perfect RedGreen skit when lifting the hawk drive in. As you are going through that sequence all I am imagining is Bill trying and failing to do something, RedGreen narrating it with a tinge of disappointment, followed by success in some form. I don't know if you planned that or not, but 8:50 my favorite part of the whole video and it's only 20 seconds long.
I would never have thought of dip switches causing issues. I learned something new from you today.
I was servicing a big graphics workstation which had intermittent failures.
I traced them to these nasty dip switches and replaced them (Approximately 15 8bit in each WS)
After that I had a quiet life with no calls from the users.
Before writing off those heads I would double-check to make sure there's no corrosion on the connectors, make sure crimps are solid, solder joints good, etc.
The cool thing is you can actually work on these drives. Newer spinning magnetic media drives are so tight tolerance wise, you really cannot work on them without specialized clean rooms and the like. Those drives were so expensive, that they HAD to be repairable.
It always gives me a bit of a startle to see what amounts to a hard drive platter just sitting there! Under almost any circumstances, if you can see that with a modern drive then the drive is _very_ dead.
my new 14TB drive is so tight tolerance that it uses helium instead of air (so the heads can float even closer I guess), its totally sealed
it made me somewhat afraid of drives, that's why I use 3-way RAID redundancy and have 8 of them. they are basically irreparable and lost data is lost forever, which is why you have to use a proper RAID system that does parity and have backups, there's no recovery if they ever fail.
which is basically the truth for any hard drive nowadays, they should never ever be used as a single drive for any system, unless you don't care about losing data, because you are going to eventually lose data.
@@monad_tcp Which RAID do you prefer? I'm not educated on the differences, and have never been able to figure it out for myself.
The low density and error checking of those makes a clean room superfluous.
1:30 WIPL stands for Warrex Initial Program Loader. There's more info in the Centurion wiki, which is googleable. You're welcome, new viewers.
Glad I got to see the mini centurion but I think your plans make a lot of sense! Nice work!
Was awesome to see it in person. Was a well done build!
Good luck! You will get there. I really like your videos and style.
Glad you are attenting VCF East! Can't wait to see what you bring with you!
Nice to see al the effort you put into this. It almost seems more simple to adapt a USB flash memory reader to the Hawk interface then mechanically reviving the unit
excellent video! good luck with the drive!
Good stuff. You have done amazing things with the old hardware and I love to watch your videos. No guarantee of success every time but always great to see you doing your level best. Carry on Please!
The thumbnail is SUPERB!
Have you been watching Curious Marc's Alto restoration videos on endless repeat? ;)
Will a Hawk drive run WITHOUT a fixed platter???
A Hawk drive shows up in the Centurion OPSYS as two logical 5MB disk drive volumes. Yes you can build up a Hawk missing one of the two platters ( fixed -or- removable ) and the heads for that missing platter can be removed. The trick is the Hawk must have sector ring pules from the fixed disk platter and the removable disk platter for the Hawk to load the heads over the remaining platter. So if you install a removable disk cart. with the platter & heads missing you still have to have the sectoring ring installed in disk cart. This is what we are now doing now, have built up a Hawk with a fixed platter & fix heads only and a cart. sector ring. Testing now under way.
I am pleased it didn't work first time. This is what makes your videos so exciting to watch.
You should try to repair broken heads. It may be broken wires to the heads because of endless bending. Check it before panic. It may be not broken coils.
I used to work on some old removable disks the video brought back some memories!! I was always careful around the heads as if we replaced one it was always a pain to get the pack alignment right. But we had only one alignment pack as they kept getting written over by mistake!!
Rule one turn ON the write protect switch anytime when using the CE Alignment Pack & rule two check rule one.
Nice work!
Fantastic effort trying to get the drive working! I’m sure you’ll get it up and running. Your tenacity and imagination will succeed in the end.
wonderful video
I followed every video of you rebuilding the Centurion. The effort you put into the system is remarkable, and the results are also very promising and educational. I have an HP-1000 minicomputer system with HP-7906 disk drive waiting to be restored. The disk drive is almost identical to the CDC one in your video, with one fixed and one removable 14-inch platters . I still have no courage to restore the drive because I have only one chance and if I fail, the data will be gone forever. Your detailed video have given me much more understanding about these devices than anywhere else. I really hope you can get the system repaired. Hopefully I'll also be able to power up my unit in the future.
Honestly with the effort and attention to detail you've put in, and the well put together narrative you weave depending on what goes down this is more entertaining to guess and hope what happens next than Yet Another Bland Marvel Movie.
I have the utmost faith you will get this working.
The head coils are likely fine, the wires that undergo flex so that the drive can seek are probably where your problem is. I had a mouse get into a Hawk and chew the head wires (this was back in the late 70s), replacing the wires did the trick, the drive worked for years after that.
Sometimes extraordinary things happen when ordinary people refuse to quit.
An absolutely excellent quote, anyone with the interest to do so can achieve stuff like this if they want it, you don't need to be a genius, building and fixing things has never been a insurmountable challenge unless a company deliberately spends lots of effort to make it as difficult as possible, ie Apple
If anyone is going to bring back this Hawk drive from the netherworld, it’s Usagi & Friends! For me as a viewer, it is at least as interesting watching all the problem solving (even if it doesn’t fit in an episode) as reaching the end goal of it working… although perhaps you could do with someone finding a box of unused CDC Hawk heads and platters!
Maybe he could get curious Marc and his crew to help. They also have experience with this kind of thing and they might also have the contacts for spare heads.
Is it posible that the leads between the heads and the connectors have failed?
@@GrayRaceCat that has been proposed in the comments many times. Guess we will find out next video. But even if they are, can they be replaced. That depends on how they are attached to the heads, might be cast in our something.
More power to ya brutha! 🤞
Very nice work indeed, I just still can't believe you are taking the hawk on a road trip :D -Looking amazing, wish I could assist more :)
You Sir, are a Legend!
2:48 Glad you're going to VCF East! You'll love it! May I also cordially invite you to VCFMW in Chicago in September...Let's see if we can get you a table near my Forgotten Machines display. The Centurion will fit in perfectly! I constantly have both real floppy/hard drives in real life and in emulation going on all of my systems. Most of them as rare and enigmatic as your Centurion! It will be a blast...
Shim the heads for the thicker disk maybe?
Is the alignment tool just hard sectoring the disk for use with the controller? Old disks needed to be synchronized from what I remember
Side thought, Should try bubble memory.. would be period correct I think…
I hope I am not the only one that sees a smiley face on the hard drive heads.
The Commodore 1541 floppy drive mechanisms made by Newtronics had a design defect where the epoxy would fail a bit and let moisture in, causing the head coils to corrode and go open. I wonder if a similar fault has happened here.
Good job my friend
I love how the camera angles make the lower head look happy and the upper head look dismayed.
Edit: OMG your bunny looks so happy at the end!
I do hope for nothing but success in the end. I am crossing my fingers that much - structural finger overload warnings are constantly tickling my brain. Following your centurion adventures proved to be the "computer-thing" i am most thrilled about since the first 3dfx voodoo boards came around.
Very interesting where we can see the fundamental workings of computer. Thank you
Remember how you had to swap the analog board earlier in the testing to get the alignment test going? Yeeeah, that thing doesn't happen to interface with the R/W heads does it? 🤔
Sounds like both the RW and Erase heads are capable of being "driven" (signal applied to, not just read from, them). So, a control board being fried is definitely something that could knacker them, I'd suspect!
Yes the only way they would be open is too much current or badly corroded enough to break the copper wires.
That's a really tough break. I hope you manage to salvage some heads from who-knows-where... Must be some out there. Good luck!
"Hawk heads are becoming really rare!"
"Yeah? How rare.."
"Rare as hen's teeth!"
* Badum-tish *
If you have a drive head like that which has crashed badly and you find that paper isn't adequate to remove material then you can pick up some .25 micron lapping film from 3M, stick it to a sheet of float glass, and use that to lap the head. I use such film to do the final pass on knives that I'm sharpening and it leaves them with a fantastic mirror finish, so it is more than up to the task for simple hard plastics. A side benefit is that this kind of film basically never deposits its abrasive, so you can be certain that it won't leave behind any grit that might cause problems. Even when using it to sharpen knives, the only reason it stops grinding effectively is when the abrasive has clogged with metal and needs to be cleared. It's also useful to cut a small swatch and adhere it to a firm foam backing to use when trying to remove marks from paint finishes, as it is extremely fine grit and tends not to leave swirls or scratches in the paint. Just remember to wet it with water or alcohol prior to use so it won't clog the abrasive too quickly.
Also, if you're curious why paper makes an acceptable abrasive: Kaolin clay. It's almost universally used in hot-press paper like printer paper to improve the color, stiffness, and surface texture of the paper. It also improves ink bonding with the paper. The fibers of the paper also function as an abrasive, but the primary abrasive is still the kaolin clay.
One thing to test is the wires between the head and the connector. The coils may be fine but the wires can be broken.
Great job, it's real ballsy to attempt to use that platter, but it's the only way. I believe in you!
13:51 the head looks freaking excited!
The heads look so happy :)
Back in the day they were a pain in the backside to fix then, and it looks like they’re going to be a pain in the backside now. Good luck!. I remember the initial seek, heads move forward at light speed “bang and break of the max distance travel stop” I was working on a newspaper “DYMO” Computer system at the time. The pressure to get it back on line was tremendous. I took them to the repair centre in Birmingham UK. What a nightmare of a unit.
Most CDC Hawk drive voice coil servo slams are caused by a failed servo feedback loop or unplugged Tach-Coil. Do agree it takes just a couple of servo slams to break the front crash stop casting.
I just love the fact that the head look like little cute faces.
Brings back bad memories of rebuilding HP 7905 and 7906 drives 40+ years ago
I live about 30 miles away from vcf east and look forward to seeing this "beast" in person.
Thank you for the great video!
I believe in you. You have hacked your way out of so many other hard problems, I think you will hack your way out of this one.
how to make head. measure ferrite permeability, cut head cross section and count turns, file down ferrite core, wound thin enameled wire, encapsulate in epoxy
Amazing! Hope you can find some working heads. Got to admit, I'd have been inclined to build some interface to make a SSD look like a Hawk drive to the machine, and put that in the mini case. Cheating, I know.
That crashed platter would make for one hell of a wall ornament
See you at VCF!
Bunny is having a lot of fun, even if you aren't!
I love how the drive heads look like little faces lol. Interesting how the drive itself can tell when you're trying to purge the insides of it.
You've come so far.
That head looked sad before cleanup and then after cleaning it looks happy :D
The enamelled copper wire used for the head windings might have been soldered using acid flux, which can eat through the joint. Worth checking if you can get to it
Acid flux wasn’t used in electronics back then. (It’s increasingly used today, as water-washable flux, but cannot be used in all situations.) Anyhow, are you perhaps thinking of corrosion from chemical stripping? It’s common to strip enameled magnet wire using chemical stripper, which is basically molten lye into which the wire ends are dipped. The wire is then rinsed repeatedly, including a neutralizing dip. Certainly if the rinsing was incomplete, it could cause corrosion down the line.
Well, I commend you for trying to fix that drive. I have worked on similar drives early in my career and those heads are pretty durable, but maybe you should consider an alternate path. I noticed you have several CDC Wren drives, I would maybe see if you could get one of those to work. And as an added benefit it would fit in your existing case. Good Luck!
People nowadays are spoiled with their fast and reliable hard drives. They need to know the pain of using floppies for everything.
Maybe someone can set up a way to load things over a serial connection. If the Centurion's serial ports are fast enough, it could be a viable alternative.
You'll figure it out eventually, man!
Strange about the heads. I'd be testing the cards carefully to make sure they're not somehow burning them out. I've not heard of that, but then I don't know anyone who's running drives that old.
Beats me but Commodore 1541 floppy drives with the Mitsumi mechanism also have head coils that fail open, I haven't found an explanation. It's also irreparable.
@@8BitNaptime If it's not being burned open, then likely some material used is corrosive over time and eating the wire open.
@@russellhltn1396 Yep that's my thoughts as well. Reminds me of the glue used on PCBs from the 80s and 90s that would turn brown and become conductive or worse corrode component legs off and wipe out traces.
@@russellhltn1396 Yes that's what makes the most sense but I don't know enough (or anything, really) about materials science and where to begin investigating. I have several dead Newtronics mechanisms to take apart the head but even if I find the cause, it's almost impossible to do anything about it, I'd say. It's awfully fine wire wound very small...
In the late 70s and early 80s I was a tech that had to deal with multiple platter CDC drives that could be removed and switched for backups etc. I was just a low level tech and the whole process of switching multi platter cartridges between 2 drives for daily backups was stressful since I never knew what was going to happen when they both spun up. A whole newspaper was in the outcome. Glad those days are over. Sometimes had to run only one drive because the other was down. No fun at all and wondering if the remaining drive would survive.
As much as I appreciate the real thing, for a mobile demonstration system I'd go with an emulated drive. The real Hawk is too big and fragile to be hauling around. -- Looking forward to seeing this at VCF East!
Please, if you haven’t already, consider putting your own spin on the recently released new old stock “Nabu” computers that were discovered from the 1980s, and that are now being sold brand new for dirt cheap. I’m about to get my own, but will need guidance on what to do with it. There are some AMAZING things being done with it already, e.g. the “ElectroGeek64” channel, but I would be fascinated to see what you can do with an original, yet brand new, 8 bit computer from the 1980s…
Great stuff. I hope to see you at VCF (if I can make it this year.) Any chance of doing a virtual drive (you know .. use a flux capacitor and a few muffler bearings to make an interface that looks just like the drive, but accesses an SD card?)
There's no fate but what you make for yourself.
Those read/write heads look like 2 smiley faces, strange enough!
You have got to figure out how to adapt more modern storage to this thing. 😅
Is it definitely the coils that are bad, and not the wires to the coils? It lookslike those flex a lot in service.
Wow... this is worth checking out. It may save you some real headaches and expense.
Startech products seem to be very well built and sturdy. I’ve been using their gear for probably about 20 years.
I wish you luck. this just seems brutal.
Oh, good… I was thinking about one of those racks for my 68010/Z80 project. 🤔
Are you sure you don't have some open diodes? I know they usually fail shorted, but I have had a bunch fail open.
make or get a couple of countersunk bolts for the wheels
You have so many brilliant contributors, maybe someone can come up with an interface for a modern mass storage device?
i like this guy ...he takes one on the chin but stood back up for round 2 :)
Where did you measure the resistance of the heads? At the heads them selves or from the end of the wires that go to the heads? Are the wires going to the heads possibly the part that's open circuit and you would just have to replace part of the wires? Likely the part of the wire that bends as the heads move.
I just want to say, the heads look like faces and are very cute.
You might want to get in contact with Shark Scrapper TH-cam channel. He's just scrapped out a Phoenix drive (floor standing model) and says he has a second one.
CDC Hawk Drive??? That's a blast from the past! I have one of these, (covered up) complete with a TI-99/10 Mini Computer in a corner of my garage! I also have 2 ModComp J1 Mini Computers. Between the two ModComp machines you could get one working computer. One machine is missing the power transformers and one machine has a bad memory core plane. The CDC hawk drive has a head platter crash so it would need a new fixed platter and new heads. I also have a set of boards for the Hawk Drive of unknown state. Any takers? I hate to relegate these to a landfill. At one time I had a whole stack of removable carts and a, "CE Alignment Disk". Now I wish I still had those to donate to the cause. I have cleaned crashed heads on these drives by using this paper/alcohol method and a microscope with the heads off of the drive.
There's some contact details on a reply to a previous comment, if you want to donate to the channel - see the reply to Eric Dravin a bit earlier...
Bummer.. coils can burn out in that older stuff (make sure there isnt a fault in any of the logic cards causing them to burn out!) have you verified the coils are actually dead and there isnt a break in the long wiring from the heads, there is quite a bit of flex in the cable and it could cause a break over the years.
I wonder if the bad analog board played any part in the heads going open circuit?
That was the first thing I thought of to.
This video was featured on Hackaday, and some folks there might have tools to align the drives, so please have a look.
13:03: Pareidolia alert!