I worked at the MPI plant in OKC where the Finch and Wren drives were made. The drive you have is actually a Wren II. I have in my possession a Wren I (36MB) and a Wren II (72MB). I used them on an Apple II believe it or not! Anyway, both are SCSi interface. Let me know if a SCSI Wren would be useful to you. BTW, do you see that round OKC test stamp in the Finch circuit board? It is black ink, and the engineers figured out it contains enough carbon to actually cause problems depending on where the test stamp is. They switched to purple ink, because it was found to be the least conductive of all the colors. Just a little trivia for you... :)
That's awesome that you used to work at MPI! The Centurion is unfortunately a bit different from the SCSI interface. As I understand it, the Wren they used was pretty much identical to the Finch and didn't even require a microcode update on the FFC controller card. Thank you so much for asking though! And thank you for the trivia, I never would have thought that the conductive properties of ink could cause issues!
@@UsagiElectric Indian ink is made from finely ground carbon black in a water based adhesive glue. That gives the black, and thus it is conductive, in that you can use it to repair remote control rubber pads that are worn. Explains why all the military boards I used had that stamp applied, after the conformal coat was applied, though the ones from early in the process, from assembly and electric test, under that coat, were normally white, and well baked onto the board as they underwent thermal cycling and drying.
I'd take a look at that voltage regulator again. Something broke the first one so it's worth checking that there isn't something causing it to fail. I don't like simply replacing blown fuses without trying to determine why they blew.
Older regulators have a tendency to die over time, they were built with quite a bit less protections than newer ones so a second of a slight overvolt or overcurrent event could easily kill one.
I always suspect capacitors when I find an inexplicably shorted 78xx or 79xx regulator. I’ve seen bypass capacitors which intermittently short, and finding those can be pretty challenging since they aren’t always shorted. I believe what I’d try next would be to get an IR camera set up with the board powered up, then look for anything inexplicably warm on the board. It is pretty unusual but not impossible for TTL ICs to spontaneously go bad; it is usually an input out of range or an overloaded output. The line driver being blown is more explainable because it is connected to an environment where there might be high electrical noise - which is by definition the reason for using amp differential driver.
@@wtmayhew Suspect one of those moulded ceramic chip capacitors is shorted, they are rather notorious for doing a flame act if the 5V rail was capable enough. Simplest test is to grab a thermal camera, and pass a 1A current through the power rails, and look for the warm traces that get cold right by the capacitor, or a warm capacitor.
Please, just FYI, we don't care how long your videos are going to be. We just care about our retro entertainment dose. Thank you so much for your videos.
On a side note... when you showed the pictures of a finch, I had to smile. I had a bullfinch as a pet a very long time ago. It learned to whistle several different songs. The poor thing was found in the rain gutter and nursed back to life. We ended up keeping him and he lived to a ripe old age.
@@WhoLover I actually have a separate channel for this! A couple videos are already up, including several interesting drives. It's "DiskMaster" with the poorly drawn profile picture
The Finch may not have lasted as long as everyone hoped, but your excitement when the Finch chirped a bit was worth the hard work everyone has put into the Centurion project.
Tip for removing ribbon cables: Cut a thin plastic sheet to the width of the cable to make a shim. Insert the plastic shim into the connector along the contact side of the ribbon cable. Grip the plastic shim and cable and gently remove. This should reduce the potential ribbon cable damage.
Sounds like a good suggestion but I think the cable in the video wasn't your typical ribbon cable, it was solid tinned copper wires stripped and pushed into individual connectors. Feels a bit scary but if there's enough slack you could actually cut and strip back the cable if everything went south.
Just playing 'armchair quarterback' here, but I'd check that 7518 voltage regulator. If the first one failed because it was overloaded, the new one may have also. If that's the case, need to find a way to check the current load on it and try to track down an overload. That's just my two bits worth. :) Regardless, also enjoy watching you revitalize this old hardware. I remember doing head alignments on CDC 300 MB disk-pack drives (size of a small washing machine) and yeah, this stuff can be fun to tinker with.
when replacing a regulator I always like to measure the output for shorts once the regulator is removed. Already saved me from burning another regulator a couple of times
I had a really old HDD that still worked for years. Until one day I tried to run it and it failed. The "rubber" seal between the two drive covers had turned to liquid! I had never seen that before. Great video!!
Dude! Your enthusiasm for this stuff is incredible. That, combined with your ability at electronics, mechanics, and electromechanics is a winning combination. The setback is just to keep you interested - how boring would it be if it was easy? 😀
Regarding the failure, the line driver was bad and the 7818 regulator was also defective after you replaced the line driver with a new one. Could it be that those components could have failed because of a root cause that is elsewhere? Because I'd look for that before burning through any more components and risking a cascade failure because the fault keeps moving around the board.
Love the enthusiasm when you were able to read data. Just goes to show, however, just how fragile these old pieces of equipment are. If it's not power it's logic, and if it's not those it could be media, although credit where credit's due, those platters and heads still appear to be in good working condition. Looking forward to the next episode.
Thank you! I actually talked about that file, but it didn't make the final cut. It was one of the first that I opened up, but the file is empty. There's nothing in there. Either it's on a bad sector or something, but there's unfortunately no data in there to save.
I was suspecting the reason for the missing ready signal on the FFC card was due to a fault on the Finch drive. Given the signal is stuck high, a short to found on the Finch (or the ribbon cable). I'd look for all possible sources of shorts to ground/power on the ready pin, adjacent wires on the cable, tracks near the ICs involved, bent pins etc.
Maybe the Finch is actually killing drivers on the FFC card? This error matches what the drive did when hooked up to the first port, Could be worth looking into?
Weren't the 'old' Wren and the ST-412 "Winchester" a very similar protocol? I looked over the documentation and it kind of appeared like the FFC used something very similar to the Shugart floppy signals with only a few pins swapped on the connector. And the 8" CDC floppy is more or less a Shugart drive with just some minor modifications on the interface connector (Adrian made one of them work on an IBM PC). So could be that the FFC was just a run of the mill floppy board with some deliberate idiosyncracies to make it incompatible with anything except CDC hardware, and those would be pretty trivial to reverse. Things like reversed pins, inverted signals or missing terminating resistors. I wouldn't be surprised if you could make an adapter cable to run standard Shugart floppies and Seagate drives on the FFC (or the other way round if you really want to).
I would also recommend getting a fan blowing across the circuit board when running outside the chassis… with all the heat generated by that voltage regulator, it’s possible it overheated and popped again. Normally it would run inside the chassis with airflow across the board helping wick away some of the heat…
Wow. When you held up the 5.25" hard drive it looked familiar to me. I actually had that drive but as a SCSI 155 meg drive which ran my BBS during my teens in the 1980s. Bought the drive used at a computer swap meet. Wow..talk about going down memory lane! 😁
I have some old 8" disk CP/M systems floating around, they have very similar looking hard disks in them. They're in storage at the moment so I don't know if they're CDC. I hope they are because your fault finding and repair will be very very useful. Thanks for putting the effort into getting these disks running.
I really miss this type of computing. it's just so boring nowdays (thought the compute power is so much nicer lol) Great job on the repair work so far and I am sure you will get this fixed again :)
I was little surprised that you were able to fix the 18V regulator by replacing it. All the symptons pointed that something was shorted on the output and that's why the regulator was heating up so much. But I guess you did some tests out side camera to confirm there were no short and obviously you got it working if even a little awhile.
@@gelo1238it wouldn't be a full short in that case I would think. But something that makes it work at too high a temperature, meaning something out of tolerance and that is a crappy fault to find. And considering the fault at the end is not the same I would think something else failed this time.
I have been a component level guy for 45 years now. It is not surprising to see a 78XX series short out like that. I may have lifted the output leg first to see if it was shorted, but I don’t really judge others techniques, because there are always mitigating factors that drive the troubleshooting process.
The older NS and Motorola voltage regulators were kind of notorious for having batches that failed with time. Generally the batches that passed the MIL testing were fine, because a failed one in any test got the entire batch canned from further use, and those instead went to the commercial market, as the MIL testing was really hard on them, and thus showed up any process deviations well. Also they tended to have diffusion carry on when run hot, meaning they eventually diffused the junctions through the die, and it went short or very leaky, resulting in it either shorting out, or going to ground. Any negative voltage transients, like an input that dropped below output too fast, also resulted in them latching up internally, and burning the pass transistor out. Especially for the higher voltage parts, had a few where it was very unhappy with more than 5V across it, as the die seems to have delaminated from the internal heat spreader, as it would go into thermal shut down around 100mA of current, or with over 10V across it.
I remember WAY back in the day - getting SOOOOo excited when I fixed something. A lot of people don't appreciate this excitement. Thanks for the video. Fun to watch.
Oh wow, I didn't realize before, but I love that digital oscilloscopes are so computerized now that you can take direct-feed screenshots from them! I remember when one or a few of the early ones I saw had a floppy disk drive, and I was like, "Oh my heck!" And now they have a USB jack and/or maybe an SD card slot instead!
What I like even better are modern ones with good web interfaces, which let you grab a screenshot from the live stream of the scope display, right from the web browser on your PC, where you’re doing the documentation.
Dude, you were right there! It's so tantalising to know so much data is just out of reach! Really awesome progress, despite this new setback. It's probably something simple like another chip gone bad. I'm sure you're already on top of it though, I can tell you're getting better at this kind of troubleshooting.
i used to operate a minicomputer in the 80s for MetLife - it was a Honeywell system that used a similar appearing removable platter "Lark" drive. I bet these are related.
"Woohoo!!! Succe...... awe shit...." Don't worry, you have the technology, you can rebuild him! At least you know the platter, heads and mechanism are good. So as long as nothing involving a special custom IC goes, you SHOULD be able to fix it given time and effort... Hopefully the other one won't be so fiddly once you figure out what's wrong with it...
Hi. I’d save the data too, if you intend to try to compile and run the source code; that way it has some test data you can use to check the functionality. If you’re just going to back up the source code to an Internet archive, then I guess the data can go bye-bye, and you can free up space on the drive. Finally, if you feed the data into a parallel-in/serial-out shift register, you could feed it into your tube computer in lieu of the paper tape drive, though you’d have to buffer it, and slow down the serial-out clock to match the data rate your tube computer can handle. 😊
if it was used for a business, it's possible some of the files contain data you don't want everyone to look at - like personnel, financial, and tax records. Even 40 year old data can potentially be abused, so it's not a good idea to just upload everything to the archives without at least reviewing the content.
can't imagine how many hours you spent on this thing and suddenly you see the files listed on screen... I know the feeling, it's amazing! Thanks for sharing that with us!
I'd guess they used this as their Repair Order/Work Order system. Generating repair orders, assigning, tracking and closing them out afterwards. A common practise in the Oil & Gas Industry.
Working on those drives must be nerve-wracking. About any and all electronics can be replaced. But platters and heads are in the list of difficult if not impossible to replace. If they fail it's over, barring luck and specialized equipment. CRT, same thing, but a little easier to find if they happened to use a common type also in TV's or something like that. I wish you luck . They one thing i would add is that i would have started on the drive with the short. Might be easy to find and it might be ok after that. A easier reference than fixing your often falling one.
This time, the "drive ready" signal died. That signal is sent by the drive to the FFC to tell the FFC that the drive is working and ready to send/receive data. So, unless the FFC is shorting the ready signal, it is most likely the drive that broke and won't sent the signal, and if it was shorting, it would likely be at some in-between level due to the voltage drop across the wire (and short), so it is about 95% certain the drive died again. Though it's also not impossible both the drive AND FFC took a dump at once, with stuff this old, you just never know.
Good stuff... really wanted /want to see more of this. When you got so very excited I was in fear you'd knock the drive off the table! Keep up the good work.
Thank you for this video. It brought back memories. When I was a grad student in 1978, I inherited a PDP-8S (serial 12 bit word) which was obsoleted by the High Energy Physics group (they had all the money!). I got the machine working, including taking apart and cleaning the hard disk, which had a whopping 52kW of storage. It had fixed heads that were a fine copper wire wrapped around a couple of (looked like) transformer laminations. No miniturisation here. I loved the way the machine casually read paper tape at (from memory, so it might be wrong) 1k characters per second, and punched it at several hundred. If we tried to make one of these today, the cost of the mechanicals would be horrendous.
I don't know much about these computers, but your excited reactions when you get stuff working makes me love watching. I can imagine theses old beasts can be a major pain to work with, so getting them to function is quite an accomplishment and deserves the reaction.
I think you're gonna end up having to replace almost every ic on that drive pcb before that trive will be anywhere near reliable again. Then again due to old age that drive my well just have mere minutes left of lifetime before the heads crash, and im willing to bet its an absolute guardian angel miracle that the heads have not crashed yet. Btw what is that flashing bunny face thing you build in the intro ? Cannot find a video about it on your channel.
I might be too much of a (retro-)gamer, but @21:50, there is a thing listed "OIXXGAMES" ... I'd be sad, if that is not interesting. Could be just a coincidence of characters...
For the other finch drive with the 5V tied to ground, a current tracer probe can be handy as it shows you where the short is by "listening" to a magnetic field generated by a pulse train injected in the 5V rail. It can be tricky to use but probably faster than troubleshoot an entire 5V rail 😄. Marc Verdiell from CuriousMarc used it sometimes. The set I know is the HP 545A (logic probe) HP 546A (logic pulser) HP 547A (current tracer). You probably need only the last two. Love the videos as always BTW!!!
My first HDD of my first computer, a 486dx, was a Quantum Fireball 850mb with windows 3.1. It was a good drive, slow and noisy but at the time we didnt knew any better. Then it started. 1 bad block, 2 bad block, 8 bad block, 40 bad block, 120 bad block. Nobody knows why, maybe too much heat, maybe the grease jammed the bearings, but the thing was dying. Some time later we bought a 20gb Samsung, it was night and day, fast and silent. Lasted for 15 years.
I totally get your enthusiasm at seeing the drive actually start to work. I was the same way this weekend when I saw the Amiga 2000 I have been restoring for the last month finally power up and ask for a workbench disk. In both cases, there is more work to do but just getting that confirmation that it's at least trying to work is a massive relief.
Just a hint: As the 7818 ate itself before, this might be an indication of something more wonky downstream. You can build in a bit of protection by lifting the output pin out of the board and put a low value resistor in series with it (on side connected to the 7818 output, the other end stuffed into the now empty PCB hole). This achieves two goals: 1 - limit the total amount of current that the downstream circuit can pull (and thereby protecting it from overheating) and 2 - being able to monitor the power draw by measuring the voltage drop across the resistor. Given that a bog standard 7818 tops out at 1 amp, a 1 ohm resistor might be reasonable. Its burden voltage will top at 1 volt which should not be too problematic on an 18 volt supply line.
Thankfully those ancient hard drives are a bit easier to repair being of DIP flavors. And note some vintage hard drives actually have their own processors (usually either 6502 or Z80), basically predating IDE technology in several ways being of early smart hard drives. And if its board is toast, it's possible to make a new board with Raspberry Pi Pico to read the data from the broken and frankenstein'd hard drive to grab whatever you want off it, it's important to find stepper motor inverter driver chip that fits the head stepper motor. I'd have to do something for MFM hard drive revival project someday as a way to rescue the data if the platters are known to be good, no scratchy noises.
hard drive revival is a pain, trying to recover one myself, a newer ide drive suffering an electrical fault... there are no companion drivees that i can compare to for thsi one unfortunately, some limited production nonretail drive so ive only been able to find a handful of listings for orphaned controller boards for this model, bought one and hoping i can use it to get the data off the drive
I hope poor kitty has had the cone of shame removed. Like AiOinc said, I do appreciate your enthusiasm. You mentioned Discord in you video. I have a 5.25" MFM hard drive project. I didn't think about using Discord to help troubleshoot it. Thanks!
Loved this video. I was smiling all the way when you revived it. You'll get it backup again since it seems like an electronical issue. Can't wait for the follow up!
iIreally enjoyed watching your detective work on the old FINCH drive. The posting made me think back about some of the old drives I've worked with. I have found that a lot of these old 8" SMD-style hard disk drives, if they haven't been physically abused, tend to still work well today. They may have issues like the line driver issue you found, but other than little things like that, they tend to just spin up and work. The first 8" drive I ever played with was years ago. It was a Micropolis 80MB drive with a proprietary high-level interface. The drive electronics and the controller that provided the high level interface were a part of the drive. It was LOUD and slow, but it worked. I had it hooked up to a Tektronix Board Bucket system with a Motorola 6809 CPU board running the Flex operating system. It was so cool to have that much storage on my homebrew computer. Back at that time (early 1980's, I seem to recall), people who had enough money might be able to buy a computer with a 5 or 10MB 5 1/4" disk drive. Having 80MB was downright huge. The very early Magnetic Peripherals WREN I drive you mentioned likely used the ST-506 interface. ST-506 is a low level interface, rather like the interface on the FINCH drive. ST-506 uses a daisy-chained command cable that goes to all of the drives, and so-called radial cables that contain the read/write/gate signals for each drive. Each drive has its own radial cable, and most controllers supported either two or four drives, each with its own radial cable. I still have a few ST-506-interface XT-1140. This was Maxtor's first product. It's an 140MB ST-506 interface drive. I have a Adaptec ST-506 to SCSI converter, so I can use this drive as a SCSI drive, which makes it easier to deal with. The drive uses a voicecoil positioner that makes all kinds of interesting noises when the drive is first powered up and it goes through its self-test seeks. It kind of sounds like R2-D2 in the Starwars movies. I think that the later WREN-series drives ended up being SASI/SCSI. I've had quite a few Magnetic Peripherals (a spin-off company from Control Data(CDC) that was mostly owned by CDC) SCSI drives from the Wren family, and still have a few of them around. All of them are SCSI -- I can't remember if they had any ESDI drives in the series or not, but I never did much with ESDI anyway. The Wren II and beyond range in size from something like 80MB up to 300MB or so. They may have actually had larger ones, but the 300MB unit was the largest one I used. The drives were rather touchy...they wanted very stable and accurate power supply voltages, and also needed very good cooling. If they got too hot, they'd start throwing errors, getting more and more of them as it heated up, until it pretty much quit reading/writing altogether. If it had good cooling, e.g., a fan blowing a good stream of air over the logic board on the drive, they would behave pretty well. I haven't powered any of them up recently, but your posting makes me want to dig them out and do some testing first to make sure the power supply pins don't short or have too low resistance to common, and then dig out my old linear hard disk power supply and make sure the gigantic caps are still good, and check to make sure it's still delivering rock-steady DC voltages, and hook it the drive up to the SCSI port on my Sun 4/280 server machine. On an aside, I love this old mid-1980's Sun Microsystems beast of a machine with it's huge VME chassis and a large and very heavy switching power supply. I think the power supply puts out something like 60 Amps at 5V along with a bunch of other voltages at pretty high amperage levels. Don't want to accidentally short that or it'll weld things together. I have a dedicated 20A circuit/outlet in my shop to power this machine. It'll heat up the space in no time flat, and makes the power meter spin way too fast, so I can't leave it running for very long. As an aside to the disk stories, the Sun 4/280 has a Fujitsu SPARC CPU running at (whoa!) 16.67MHz, and can handle up to 128MB of RAM. The system has two 8" Fujitsu SMD hard disks, similar in physical size to your FINCH drives, but they each hold 300MB. They draw a lot of power, and the head positioner in these things has so much torque that when they do their power-up seek testing, it sounds like a 9mm fully-automatic machine gun firing in full-auto mode nearby; the heads move with great authority and do so very quickly. These drives also sound like a small jet engine as they spin up. Both of the vintage Fujitsu 8" drives still work great. The Sun 4/280 has the SMD disk interface card, a Pertec Mag-Tape Interface card (to which a Sun-branded Fujitsu 9-track mag tape drive is connected), a SCSI interface card, an ALM (async line multiplexer) card with 16 RS-232 serial ports, 96MB of RAM (a 64MB board and a 32MB Board), and the Sunrise-class CPU board with 10Mb Ethernet (AUI connector), two RS-232 serial ports, and a basic monochrome framebuffer. It runs SunOS 4.1.3, which is kind of like Berkeley 4.2 Unix with some SysV stuff thrown in. Like the WREN drives, these Fujitsu 8" drives need a lot of cooling. I've got two big high-volume fans, one each blowing over each drive. The drives are very heavy. The kinetic energy contained in the drive with all of its platters spinning (thick aluminum platters with the magnetic coating hooked to a really sturdy hub to support the platters, with the motor embedded in the hub). There's a lot of spinning mass. I think the drive spins at 7200 RPM, which is part of the reason it has the capacity it does, as well as why it sounds so mean as it spins up. If a platter somehow got loose and ended up rattling around in the HDA while the drive was spun up, I'd guess that the amount of energy released would be rather scary, and the result would be total destruction of the drive. The HDA housing is made of a very thick machined alloy case (not a casting)..I'd say it's at least 3/8" thick. I think they did this to keep pieces from coming out at potentially deadly rates of speed if something bad happened inside. I hope you can get the other Finch drive working and see if there is any interesting stuff on it. Hopefully the issue with the drive you got running that died as you were trying to copy stuff off it can be resurrected without too much fuss. I love your channel -- your enthusiasm is totally contagious, and the stuff you find to work on is always interesting. Your production quality is consistently good, and I always love looking at the bunnies or kitties at the end of each episode. I can't wait until you dig into that Litton machine with the drum memory. That'll be really interesting to see. I played with an old drum memory based computer back in high school that was made by 3M in 1964..yeah, the scotch tape(tm) people. It had 8K of 24-bit words, and was a bit-serial architecture. It was all-transistor. It was built as a process control computer. Had to hand optimize the machine code to account for the drum's rotation in order to get the best performance out of it. On average, though, it could barely keep the console terminal, a Teletype 33-ASR, running at full 10 character-per-second speed. It was a blast to tinker around with, though. I learned a lot from playing with it. The only way to do indexing was to use self-modifying code, which was quite interesting, as if you wanted the indexing to run quickly, you had to structure the data you were indexing so as to account for the rotation of the drum. You had to do effective address calculation in your code to generate a "new" instruction to access the next data item and then store the generated instruction in its place, where it'd be executed to fetch the data. You didn't /have/ to optimize the code, but if you wanted the best performance out of it you had to do the address calculations for the data the instruction was to operate on, as well as for the address of the next instruction to execute. It was quite a simple machine, with an accumulator and a scratch register called the B register. The B register was used for all I/O data, and could be used as a scratch register for temporary storage of intermediate results of math calculations. As I usually do, I rambled on way too long. Hopefully someone finds the reading interesting. Good luck with your continued work on the Tube Computer, the Bendix G-15 (another one I'm drooling over, and can't wait to see more videos on it), the LItton Computer, and of course, the really cool and rare Centurion machines. God bless. Rick Bensene, The Old Calculator Museum
Thank you, great video. I have a Wren 6 (~330MB) running in an AT and I never knew what the 'wren' part meant. I've seen your videos on the finch drive and never made the connection until now!
Check that 7818 temp again, maybe it died because something downstream is drawing too much current. I have a similar desoldering iron. I found it in a recycle bin and it worked fine, one of the best finds ever.
I love it that you effectively got all equipment working - even only for a little bit - that you started this Centurion series with. Or there anything that you haven’t gotten working yet?
Yup, got the printer and the phoenix drive left! I also have a 128k MEM card and a MUX card that have stopped working. The current plan looks like this: Repair FFC, MUX and MEM cards Repair both Finch drives Restore printer Restore Phoenix Design and build a paper tape card and drive from scratch Lots of work to go, but we've managed to work our way through the good majority of stuff that came with the machine!
When a device suddenly fails after being powered on for a while, I think heat build up is causing a component to fail. Does a cold start after a period of rest show the same problem?
1:46 those are the samw size platters used in the old Prowler. I remember back in the late 90's at NAS Whidbey the Avionics techs would bring us crashed platters and we would media blast them removing the magnetic media (analog bleach bit)
The sound of that disk spinning up reminded my of my old ICL Perq workstation that had a similar drive. You've shown it spin up, but how long does it take to spin down? I remember the floor shaking a few minutes after the Perq had been powered down.
My bunny! After Usagi Yojimbo you are my favourite usagi of all time! I saw a 'games???' directory on there....I do hope that comes in for some early exploring once you have it restored again, as long as it dosen't ask you to play a nice game of chess....
Thanks! That was actually one of the first files I opened up! I was really curious what was inside, but unfortunately the file is empty. There's nothing in there. Either it's on a bad sector or was something they were planning to write data to and never got to, but sadly it's just totally blank.
@@UsagiElectric Ah, what a shame. Glad you checked it before the drive died though, it would have driven me (and certainly you) nuts should the drive prove unfixable a second time!
This brought back a lot (A LOT) of memories diagnosing PC problems in the heydays of the 80s/90s...nothing has ever been quite like this era of technology. I would often spend many hours researching or figuring out an issue, digging into the most esoteric of possibile causes... only to fix the problem, have it boot & run successfully & then develop a new problem about 30 minutes later 😅
I mean, it hasn't changed that much. Just the information's gotten a bit easier to access. Software issues especially sometimes make me feel like I'm playing Whack-A-Mole.
To dump the drive, it would be interesting to run a read test (or make a "dd" like programme) but then actually tap all the signals of the drive into a logic analyzer/scope and share the waveform of the complete run. It would provide not only raw data, but checksums and low level geometry (as we can see the index signal).
Always nice to end the Sunday with a Usagi Electric video! Meh, too bad the Finch 'died'. I'm still very impressed with your fault finding and the progress you already made. Keep going, we love it!
The following is from Chat GPT.... "Here are a few examples of capacities for different models within the CDC Wren series: The CDC Wren I had a capacity ranging from 63.5MB to 340MB, depending on the specific model. CDC Wren II: The Wren II, released in the late 1980s, offered capacities ranging from 130MB to 330MB. This was a significant leap in storage capacity compared to earlier hard drives. CDC Wren III: The Wren III models, released around the same time, featured capacities that typically ranged from 540MB to 1.2GB."
Ha... I, too, have a legacy 8" drive. It is a Hitachi 168MB. Weighs in at about 35#. It never ran. Cost me $3300 in 1981. It's current use is to weigh down out picnic umbrella.
I worked at the MPI plant in OKC where the Finch and Wren drives were made. The drive you have is actually a Wren II. I have in my possession a Wren I (36MB) and a Wren II (72MB). I used them on an Apple II believe it or not! Anyway, both are SCSi interface. Let me know if a SCSI Wren would be useful to you. BTW, do you see that round OKC test stamp in the Finch circuit board? It is black ink, and the engineers figured out it contains enough carbon to actually cause problems depending on where the test stamp is. They switched to purple ink, because it was found to be the least conductive of all the colors. Just a little trivia for you... :)
That's awesome that you used to work at MPI!
The Centurion is unfortunately a bit different from the SCSI interface. As I understand it, the Wren they used was pretty much identical to the Finch and didn't even require a microcode update on the FFC controller card. Thank you so much for asking though!
And thank you for the trivia, I never would have thought that the conductive properties of ink could cause issues!
How ironic. "Tests passed, stamp it OK! Oops, now it doesn't work any more..."
@@UsagiElectric Indian ink is made from finely ground carbon black in a water based adhesive glue. That gives the black, and thus it is conductive, in that you can use it to repair remote control rubber pads that are worn. Explains why all the military boards I used had that stamp applied, after the conformal coat was applied, though the ones from early in the process, from assembly and electric test, under that coat, were normally white, and well baked onto the board as they underwent thermal cycling and drying.
I'd take a look at that voltage regulator again. Something broke the first one so it's worth checking that there isn't something causing it to fail. I don't like simply replacing blown fuses without trying to determine why they blew.
Older regulators have a tendency to die over time, they were built with quite a bit less protections than newer ones so a second of a slight overvolt or overcurrent event could easily kill one.
I was surprised that there was no check of the output pin of the regulator to see if there was a short.
I always suspect capacitors when I find an inexplicably shorted 78xx or 79xx regulator. I’ve seen bypass capacitors which intermittently short, and finding those can be pretty challenging since they aren’t always shorted. I believe what I’d try next would be to get an IR camera set up with the board powered up, then look for anything inexplicably warm on the board. It is pretty unusual but not impossible for TTL ICs to spontaneously go bad; it is usually an input out of range or an overloaded output. The line driver being blown is more explainable because it is connected to an environment where there might be high electrical noise - which is by definition the reason for using amp differential driver.
@@wtmayhew Suspect one of those moulded ceramic chip capacitors is shorted, they are rather notorious for doing a flame act if the 5V rail was capable enough. Simplest test is to grab a thermal camera, and pass a 1A current through the power rails, and look for the warm traces that get cold right by the capacitor, or a warm capacitor.
Could be input overvoltage (ac ripple?) or bad thermal contact (use paste)
Please, just FYI, we don't care how long your videos are going to be. We just care about our retro entertainment dose. Thank you so much for your videos.
Well spoken. I can only concur regarding the retro dose. 😊
Yeah, but you don't want to go above 30 mins, as the HDD cannot run past that time...!
On a side note... when you showed the pictures of a finch, I had to smile. I had a bullfinch as a pet a very long time ago. It learned to whistle several different songs. The poor thing was found in the rain gutter and nursed back to life. We ended up keeping him and he lived to a ripe old age.
As a collector of vintage hard disks, I appreciate your enthusiasm
Would you want to show them off on your channel? I've subbed and turned on notifications regardless
@@WhoLover I actually have a separate channel for this! A couple videos are already up, including several interesting drives. It's "DiskMaster" with the poorly drawn profile picture
This here is that channel, if you like.
@@TheDiskMaster oh, excellent! Just subbed :)
Any idea what interface a discus brand drive uses? What kind of controller? Mine has a 50 and a 20 pin connector.
The Finch may not have lasted as long as everyone hoped, but your excitement when the Finch chirped a bit was worth the hard work everyone has put into the Centurion project.
Tip for removing ribbon cables: Cut a thin plastic sheet to the width of the cable to make a shim. Insert the plastic shim into the connector along the contact side of the ribbon cable. Grip the plastic shim and cable and gently remove. This should reduce the potential ribbon cable damage.
Sounds like a good suggestion but I think the cable in the video wasn't your typical ribbon cable, it was solid tinned copper wires stripped and pushed into individual connectors. Feels a bit scary but if there's enough slack you could actually cut and strip back the cable if everything went south.
I do not know half of what you are talking about but i enjoy watching somebody who never gets frustrated. Great stuff.
Just playing 'armchair quarterback' here, but I'd check that 7518 voltage regulator. If the first one failed because it was overloaded, the new one may have also. If that's the case, need to find a way to check the current load on it and try to track down an overload. That's just my two bits worth. :)
Regardless, also enjoy watching you revitalize this old hardware. I remember doing head alignments on CDC 300 MB disk-pack drives (size of a small washing machine) and yeah, this stuff can be fun to tinker with.
that is what I am thinking - resistor gone low ?
Ha! Love your enthusiasm - well deserved! The success blew my mind. It fills me with determination.
Thank you so much Keri!
when replacing a regulator I always like to measure the output for shorts once the regulator is removed. Already saved me from burning another regulator a couple of times
I concur :)
The index clock of 60Hz seems perfect: 60 revolutions per second are 3600 revolutions per minute, a standard speed for hard drives back then.
That makes sense! My brain wasn't good enough to make the mental connection while the camera was rolling, haha.
Oh wow, now I get the 5400 and 7200 more modern speeds.. And they're all multiples of 300, so fine in either 60 or 50Hz realms
Always a good day when you upload a new video! The look on your face and the celebration when you saw that directory listing was priceless!!
I thought the old drive was well and truly dead, and then bam, life! It was an awesome moment!
I had a really old HDD that still worked for years. Until one day I tried to run it and it failed. The "rubber" seal between the two drive covers had turned to liquid! I had never seen that before. Great video!!
22:13: "It freakin' works!"
Channeling your inner Adrian Black, I see :-)
Dude! Your enthusiasm for this stuff is incredible. That, combined with your ability at electronics, mechanics, and electromechanics is a winning combination.
The setback is just to keep you interested - how boring would it be if it was easy? 😀
Haha, yeah.
In the 90's I used to work at the Seagate plant in OKC and worked on the Wren III line in testing. That was a fun job.
Regarding the failure, the line driver was bad and the 7818 regulator was also defective after you replaced the line driver with a new one. Could it be that those components could have failed because of a root cause that is elsewhere? Because I'd look for that before burning through any more components and risking a cascade failure because the fault keeps moving around the board.
Love the enthusiasm when you were able to read data. Just goes to show, however, just how fragile these old pieces of equipment are. If it's not power it's logic, and if it's not those it could be media, although credit where credit's due, those platters and heads still appear to be in good working condition. Looking forward to the next episode.
Outstanding work. Just like the last controller "fix"! So curious about OIXXGAMES
I noticed that too, at 21:52
First thing I noticed as well, and then he got started looking at the sources. Different priorities I guess 😂
@@johnm2012 LOL me3 (or4!)
Thank you!
I actually talked about that file, but it didn't make the final cut.
It was one of the first that I opened up, but the file is empty. There's nothing in there. Either it's on a bad sector or something, but there's unfortunately no data in there to save.
I was suspecting the reason for the missing ready signal on the FFC card was due to a fault on the Finch drive. Given the signal is stuck high, a short to found on the Finch (or the ribbon cable).
I'd look for all possible sources of shorts to ground/power on the ready pin, adjacent wires on the cable, tracks near the ICs involved, bent pins etc.
Maybe the Finch is actually killing drivers on the FFC card? This error matches what the drive did when hooked up to the first port, Could be worth looking into?
This is the only channel I actual like the intro to. Short and catchy. Thanks for another awesome video!
Weren't the 'old' Wren and the ST-412 "Winchester" a very similar protocol? I looked over the documentation and it kind of appeared like the FFC used something very similar to the Shugart floppy signals with only a few pins swapped on the connector. And the 8" CDC floppy is more or less a Shugart drive with just some minor modifications on the interface connector (Adrian made one of them work on an IBM PC). So could be that the FFC was just a run of the mill floppy board with some deliberate idiosyncracies to make it incompatible with anything except CDC hardware, and those would be pretty trivial to reverse. Things like reversed pins, inverted signals or missing terminating resistors.
I wouldn't be surprised if you could make an adapter cable to run standard Shugart floppies and Seagate drives on the FFC (or the other way round if you really want to).
I would also recommend getting a fan blowing across the circuit board when running outside the chassis… with all the heat generated by that voltage regulator, it’s possible it overheated and popped again. Normally it would run inside the chassis with airflow across the board helping wick away some of the heat…
Wow. When you held up the 5.25" hard drive it looked familiar to me. I actually had that drive but as a SCSI 155 meg drive which ran my BBS during my teens in the 1980s. Bought the drive used at a computer swap meet. Wow..talk about going down memory lane! 😁
Great thinking Meisaka. 👍
I have some old 8" disk CP/M systems floating around, they have very similar looking hard disks in them. They're in storage at the moment so I don't know if they're CDC. I hope they are because your fault finding and repair will be very very useful. Thanks for putting the effort into getting these disks running.
I really miss this type of computing. it's just so boring nowdays (thought the compute power is so much nicer lol)
Great job on the repair work so far and I am sure you will get this fixed again :)
Did you test the old SN75110 chip outside the drive controller? I bet you could breadboard up a test jig.
I was little surprised that you were able to fix the 18V regulator by replacing it. All the symptons pointed that something was shorted on the output and that's why the regulator was heating up so much. But I guess you did some tests out side camera to confirm there were no short and obviously you got it working if even a little awhile.
Im curious if there is still a short. Maybe regulator was working too hard and died
My first guess was a short as well. But it's an easy component to replace and they do fail.
@@gelo1238it wouldn't be a full short in that case I would think. But something that makes it work at too high a temperature, meaning something out of tolerance and that is a crappy fault to find.
And considering the fault at the end is not the same I would think something else failed this time.
I have been a component level guy for 45 years now. It is not surprising to see a 78XX series short out like that. I may have lifted the output leg first to see if it was shorted, but I don’t really judge others techniques, because there are always mitigating factors that drive the troubleshooting process.
The older NS and Motorola voltage regulators were kind of notorious for having batches that failed with time. Generally the batches that passed the MIL testing were fine, because a failed one in any test got the entire batch canned from further use, and those instead went to the commercial market, as the MIL testing was really hard on them, and thus showed up any process deviations well. Also they tended to have diffusion carry on when run hot, meaning they eventually diffused the junctions through the die, and it went short or very leaky, resulting in it either shorting out, or going to ground. Any negative voltage transients, like an input that dropped below output too fast, also resulted in them latching up internally, and burning the pass transistor out. Especially for the higher voltage parts, had a few where it was very unhappy with more than 5V across it, as the die seems to have delaminated from the internal heat spreader, as it would go into thermal shut down around 100mA of current, or with over 10V across it.
I remember WAY back in the day - getting SOOOOo excited when I fixed something. A lot of people don't appreciate this excitement. Thanks for the video. Fun to watch.
It is always when you are happy something gonna knock you down, But you never give up anyway. love the dedication!
Oh wow, I didn't realize before, but I love that digital oscilloscopes are so computerized now that you can take direct-feed screenshots from them! I remember when one or a few of the early ones I saw had a floppy disk drive, and I was like, "Oh my heck!" And now they have a USB jack and/or maybe an SD card slot instead!
What I like even better are modern ones with good web interfaces, which let you grab a screenshot from the live stream of the scope display, right from the web browser on your PC, where you’re doing the documentation.
Channeling Adrian Black with the IT FREAKIN' WORKS!
Your passion for retro tech is refreshing. Thank you for all that you do.
Dude, you were right there! It's so tantalising to know so much data is just out of reach! Really awesome progress, despite this new setback. It's probably something simple like another chip gone bad. I'm sure you're already on top of it though, I can tell you're getting better at this kind of troubleshooting.
Congrats on the progress. Only suggestion is make the obscured information more blurry, some of it is still legible.
I look forward seeing you fix the first port on the controller card as well. We can't leave anything broken ;-)
Oh absolutely!
I actually have a MUX card and a 128k MEM card that have stopped working as well, so we'll tackle all three at the same time!
@@UsagiElectric It's suprising how many faults your hardware is developing. A curse for you, a gift for your viewers ;-)
i used to operate a minicomputer in the 80s for MetLife - it was a Honeywell system that used a similar appearing removable platter "Lark" drive. I bet these are related.
It is. I searched for it on Google and I found a specs sheet for that drive by CDC
that's amazing that it was working (for a little while at least)! looking forward to the next installment
Man what a journey! I absolutely love it and it's so awesome to have these ppl in the discord!
Thank you!
The project absolutely wouldn't be where it is without the help of the people over on the Discord, they've been amazing!
"Woohoo!!! Succe...... awe shit...." Don't worry, you have the technology, you can rebuild him! At least you know the platter, heads and mechanism are good. So as long as nothing involving a special custom IC goes, you SHOULD be able to fix it given time and effort... Hopefully the other one won't be so fiddly once you figure out what's wrong with it...
Hi. I’d save the data too, if you intend to try to compile and run the source code; that way it has some test data you can use to check the functionality. If you’re just going to back up the source code to an Internet archive, then I guess the data can go bye-bye, and you can free up space on the drive. Finally, if you feed the data into a parallel-in/serial-out shift register, you could feed it into your tube computer in lieu of the paper tape drive, though you’d have to buffer it, and slow down the serial-out clock to match the data rate your tube computer can handle. 😊
if it was used for a business, it's possible some of the files contain data you don't want everyone to look at - like personnel, financial, and tax records. Even 40 year old data can potentially be abused, so it's not a good idea to just upload everything to the archives without at least reviewing the content.
can't imagine how many hours you spent on this thing and suddenly you see the files listed on screen... I know the feeling, it's amazing! Thanks for sharing that with us!
I'd guess they used this as their Repair Order/Work Order system. Generating repair orders, assigning, tracking and closing them out afterwards. A common practise in the Oil & Gas Industry.
Working on those drives must be nerve-wracking. About any and all electronics can be replaced. But platters and heads are in the list of difficult if not impossible to replace. If they fail it's over, barring luck and specialized equipment.
CRT, same thing, but a little easier to find if they happened to use a common type also in TV's or something like that.
I wish you luck .
They one thing i would add is that i would have started on the drive with the short. Might be easy to find and it might be ok after that. A easier reference than fixing your often falling one.
Cool project, man! Thank you for working to preserve our digital history!
Your energetic exuberant enthusiasm David, is very infectious ......
Are you sure the drive died and not the FFC port (again)?
This time, the "drive ready" signal died. That signal is sent by the drive to the FFC to tell the FFC that the drive is working and ready to send/receive data. So, unless the FFC is shorting the ready signal, it is most likely the drive that broke and won't sent the signal, and if it was shorting, it would likely be at some in-between level due to the voltage drop across the wire (and short), so it is about 95% certain the drive died again. Though it's also not impossible both the drive AND FFC took a dump at once, with stuff this old, you just never know.
Good stuff... really wanted /want to see more of this. When you got so very excited I was in fear you'd knock the drive off the table! Keep up the good work.
Yeah I'm so nervous about that drive, would be great to see a simple wooden frame around it to prevent a fall :)
David!!! Such a momentous win!! Congratulations sir!
Love to see my favorite Digital Archaeologist at work. What a fascinating drive. Keep up the good work!
Thank you for this video. It brought back memories. When I was a grad student in 1978, I inherited a PDP-8S (serial 12 bit word) which was obsoleted by the High Energy Physics group (they had all the money!). I got the machine working, including taking apart and cleaning the hard disk, which had a whopping 52kW of storage. It had fixed heads that were a fine copper wire wrapped around a couple of (looked like) transformer laminations. No miniturisation here.
I loved the way the machine casually read paper tape at (from memory, so it might be wrong) 1k characters per second, and punched it at several hundred. If we tried to make one of these today, the cost of the mechanicals would be horrendous.
Cool, I'm glad you got that drive working and found the data on its disk!
With some PC hard drives from this era, it was not always possible to run a drive on its side if it had been formatted while in a horizontal position.
This is like modern archeology. So captivating! 🤓
🎉🎉 well done, I’m really impressed with your fault finding
I don't know much about these computers, but your excited reactions when you get stuff working makes me love watching. I can imagine theses old beasts can be a major pain to work with, so getting them to function is quite an accomplishment and deserves the reaction.
Is there a story behind why the CDC Phoenix drive is named that? 😻
I think you're gonna end up having to replace almost every ic on that drive pcb before that trive will be anywhere near reliable again. Then again due to old age that drive my well just have mere minutes left of lifetime before the heads crash, and im willing to bet its an absolute guardian angel miracle that the heads have not crashed yet.
Btw what is that flashing bunny face thing you build in the intro ? Cannot find a video about it on your channel.
I might be too much of a (retro-)gamer, but @21:50, there is a thing listed "OIXXGAMES" ... I'd be sad, if that is not interesting. Could be just a coincidence of characters...
For the other finch drive with the 5V tied to ground, a current tracer probe can be handy as it shows you where the short is by "listening" to a magnetic field generated by a pulse train injected in the 5V rail. It can be tricky to use but probably faster than troubleshoot an entire 5V rail 😄.
Marc Verdiell from CuriousMarc used it sometimes. The set I know is the HP 545A (logic probe) HP 546A (logic pulser) HP 547A (current tracer). You probably need only the last two.
Love the videos as always BTW!!!
My first HDD of my first computer, a 486dx, was a Quantum Fireball 850mb with windows 3.1. It was a good drive, slow and noisy but at the time we didnt knew any better. Then it started. 1 bad block, 2 bad block, 8 bad block, 40 bad block, 120 bad block. Nobody knows why, maybe too much heat, maybe the grease jammed the bearings, but the thing was dying. Some time later we bought a 20gb Samsung, it was night and day, fast and silent. Lasted for 15 years.
I looks like you are going to need a few new chips to replace the burned out chips.
I totally get your enthusiasm at seeing the drive actually start to work. I was the same way this weekend when I saw the Amiga 2000 I have been restoring for the last month finally power up and ask for a workbench disk. In both cases, there is more work to do but just getting that confirmation that it's at least trying to work is a massive relief.
Just a hint: As the 7818 ate itself before, this might be an indication of something more wonky downstream. You can build in a bit of protection by lifting the output pin out of the board and put a low value resistor in series with it (on side connected to the 7818 output, the other end stuffed into the now empty PCB hole). This achieves two goals: 1 - limit the total amount of current that the downstream circuit can pull (and thereby protecting it from overheating) and 2 - being able to monitor the power draw by measuring the voltage drop across the resistor. Given that a bog standard 7818 tops out at 1 amp, a 1 ohm resistor might be reasonable. Its burden voltage will top at 1 volt which should not be too problematic on an 18 volt supply line.
Thankfully those ancient hard drives are a bit easier to repair being of DIP flavors. And note some vintage hard drives actually have their own processors (usually either 6502 or Z80), basically predating IDE technology in several ways being of early smart hard drives. And if its board is toast, it's possible to make a new board with Raspberry Pi Pico to read the data from the broken and frankenstein'd hard drive to grab whatever you want off it, it's important to find stepper motor inverter driver chip that fits the head stepper motor. I'd have to do something for MFM hard drive revival project someday as a way to rescue the data if the platters are known to be good, no scratchy noises.
you say there's nothing exciting on that drive, i was pretty excited by the date codes, this bad boy was still in service in to the 90s!
hard drive revival is a pain, trying to recover one myself, a newer ide drive suffering an electrical fault... there are no companion drivees that i can compare to for thsi one unfortunately, some limited production nonretail drive so ive only been able to find a handful of listings for orphaned controller boards for this model, bought one and hoping i can use it to get the data off the drive
I hope poor kitty has had the cone of shame removed.
Like AiOinc said, I do appreciate your enthusiasm. You mentioned Discord in you video. I have a 5.25" MFM hard drive project. I didn't think about using Discord to help troubleshoot it. Thanks!
incidentally just noticed that seagate has started shipping their 30TB+ HAMR drives
Loved this video. I was smiling all the way when you revived it. You'll get it backup again since it seems like an electronical issue. Can't wait for the follow up!
this is very underrated. thank you for your work.
Thank you! This video is a good illustration of methodical trouble shooting. Thanks for sharing.
iIreally enjoyed watching your detective work on the old FINCH drive. The posting made me think back about some of the old drives I've worked with.
I have found that a lot of these old 8" SMD-style hard disk drives, if they haven't been physically abused, tend to still work well today. They may have issues like the line driver issue you found, but other than little things like that, they tend to just spin up and work.
The first 8" drive I ever played with was years ago. It was a Micropolis 80MB drive with a proprietary high-level interface. The drive electronics and the controller that provided the high level interface were a part of the drive. It was LOUD and slow, but it worked. I had it hooked up to a Tektronix Board Bucket system with a Motorola 6809 CPU board running the Flex operating system. It was so cool to have that much storage on my homebrew computer. Back at that time (early 1980's, I seem to recall), people who had enough money might be able to buy a computer with a 5 or 10MB 5 1/4" disk drive. Having 80MB was downright huge.
The very early Magnetic Peripherals WREN I drive you mentioned likely used the ST-506 interface. ST-506 is a low level interface, rather like the interface on the FINCH drive. ST-506 uses a daisy-chained command cable that goes to all of the drives, and so-called radial cables that contain the read/write/gate signals for each drive. Each drive has its own radial cable, and most controllers supported either two or four drives, each with its own radial cable. I still have a few ST-506-interface XT-1140. This was Maxtor's first product. It's an 140MB ST-506 interface drive. I have a Adaptec ST-506 to SCSI converter, so I can use this drive as a SCSI drive, which makes it easier to deal with. The drive uses a voicecoil positioner that makes all kinds of interesting noises when the drive is first powered up and it goes through its self-test seeks. It kind of sounds like R2-D2 in the Starwars movies.
I think that the later WREN-series drives ended up being SASI/SCSI. I've had quite a few Magnetic Peripherals (a spin-off company from Control Data(CDC) that was mostly owned by CDC) SCSI drives from the Wren family, and still have a few of them around. All of them are SCSI -- I can't remember if they had any ESDI drives in the series or not, but I never did much with ESDI anyway. The Wren II and beyond range in size from something like 80MB up to 300MB or so. They may have actually had larger ones, but the 300MB unit was the largest one I used.
The drives were rather touchy...they wanted very stable and accurate power supply voltages, and also needed very good cooling. If they got too hot, they'd start throwing errors, getting more and more of them as it heated up, until it pretty much quit reading/writing altogether. If it had good cooling, e.g., a fan blowing a good stream of air over the logic board on the drive, they would behave pretty well. I haven't powered any of them up recently, but your posting makes me want to dig them out and do some testing first to make sure the power supply pins don't short or have too low resistance to common, and then dig out my old linear hard disk power supply and make sure the gigantic caps are still good, and check to make sure it's still delivering rock-steady DC voltages, and hook it the drive up to the SCSI port on my Sun 4/280 server machine.
On an aside, I love this old mid-1980's Sun Microsystems beast of a machine with it's huge VME chassis and a large and very heavy switching power supply. I think the power supply puts out something like 60 Amps at 5V along with a bunch of other voltages at pretty high amperage levels. Don't want to accidentally short that or it'll weld things together. I have a dedicated 20A circuit/outlet in my shop to power this machine. It'll heat up the space in no time flat, and makes the power meter spin way too fast, so I can't leave it running for very long.
As an aside to the disk stories, the Sun 4/280 has a Fujitsu SPARC CPU running at (whoa!) 16.67MHz, and can handle up to 128MB of RAM. The system has two 8" Fujitsu SMD hard disks, similar in physical size to your FINCH drives, but they each hold 300MB. They draw a lot of power, and the head positioner in these things has so much torque that when they do their power-up seek testing, it sounds like a 9mm fully-automatic machine gun firing in full-auto mode nearby; the heads move with great authority and do so very quickly. These drives also sound like a small jet engine as they spin up. Both of the vintage Fujitsu 8" drives still work great. The Sun 4/280 has the SMD disk interface card, a Pertec Mag-Tape Interface card (to which a Sun-branded Fujitsu 9-track mag tape drive is connected), a SCSI interface card, an ALM (async line multiplexer) card with 16 RS-232 serial ports, 96MB of RAM (a 64MB board and a 32MB Board), and the Sunrise-class CPU board with 10Mb Ethernet (AUI connector), two RS-232 serial ports, and a basic monochrome framebuffer. It runs SunOS 4.1.3, which is kind of like Berkeley 4.2 Unix with some SysV stuff thrown in.
Like the WREN drives, these Fujitsu 8" drives need a lot of cooling. I've got two big high-volume fans, one each blowing over each drive. The drives are very heavy. The kinetic energy contained in the drive with all of its platters spinning (thick aluminum platters with the magnetic coating hooked to a really sturdy hub to support the platters, with the motor embedded in the hub). There's a lot of spinning mass. I think the drive spins at 7200 RPM, which is part of the reason it has the capacity it does, as well as why it sounds so mean as it spins up. If a platter somehow got loose and ended up rattling around in the HDA while the drive was spun up, I'd guess that the amount of energy released would be rather scary, and the result would be total destruction of the drive. The HDA housing is made of a very thick machined alloy case (not a casting)..I'd say it's at least 3/8" thick. I think they did this to keep pieces from coming out at potentially deadly rates of speed if something bad happened inside.
I hope you can get the other Finch drive working and see if there is any interesting stuff on it. Hopefully the issue with the drive you got running that died as you were trying to copy stuff off it can be resurrected without too much fuss. I love your channel -- your enthusiasm is totally contagious, and the stuff you find to work on is always interesting. Your production quality is consistently good, and I always love looking at the bunnies or kitties at the end of each episode. I can't wait until you dig into that Litton machine with the drum memory. That'll be really interesting to see.
I played with an old drum memory based computer back in high school that was made by 3M in 1964..yeah, the scotch tape(tm) people. It had 8K of 24-bit words, and was a bit-serial architecture. It was all-transistor. It was built as a process control computer. Had to hand optimize the machine code to account for the drum's rotation in order to get the best performance out of it. On average, though, it could barely keep the console terminal, a Teletype 33-ASR, running at full 10 character-per-second speed. It was a blast to tinker around with, though. I learned a lot from playing with it. The only way to do indexing was to use self-modifying code, which was quite interesting, as if you wanted the indexing to run quickly, you had to structure the data you were indexing so as to account for the rotation of the drum. You had to do effective address calculation in your code to generate a "new" instruction to access the next data item and then store the generated instruction in its place, where it'd be executed to fetch the data. You didn't /have/ to optimize the code, but if you wanted the best performance out of it you had to do the address calculations for the data the instruction was to operate on, as well as for the address of the next instruction to execute. It was quite a simple machine, with an accumulator and a scratch register called the B register. The B register was used for all I/O data, and could be used as a scratch register for temporary storage of intermediate results of math calculations.
As I usually do, I rambled on way too long. Hopefully someone finds the reading interesting. Good luck with your continued work on the Tube Computer, the Bendix G-15 (another one I'm drooling over, and can't wait to see more videos on it), the LItton Computer, and of course, the really cool and rare Centurion machines. God bless. Rick Bensene, The Old Calculator Museum
I have absolutely no idea what all this was about. I just enjoy seeing the enthusiasm.
Thank you, great video. I have a Wren 6 (~330MB) running in an AT and I never knew what the 'wren' part meant. I've seen your videos on the finch drive and never made the connection until now!
Check that 7818 temp again, maybe it died because something downstream is drawing too much current. I have a similar desoldering iron. I found it in a recycle bin and it worked fine, one of the best finds ever.
I love it that you effectively got all equipment working - even only for a little bit - that you started this Centurion series with. Or there anything that you haven’t gotten working yet?
Just the printer to go
@@TheErador As well as the Phoenix drive.
Yup, got the printer and the phoenix drive left! I also have a 128k MEM card and a MUX card that have stopped working.
The current plan looks like this:
Repair FFC, MUX and MEM cards
Repair both Finch drives
Restore printer
Restore Phoenix
Design and build a paper tape card and drive from scratch
Lots of work to go, but we've managed to work our way through the good majority of stuff that came with the machine!
@@UsagiElectric thanks for sharing!
@@UsagiElectric I'd forgotten about the phœnix. Was that in the original haul or with the counterfeit?
When a device suddenly fails after being powered on for a while, I think heat build up is causing a component to fail. Does a cold start after a period of rest show the same problem?
1:46 those are the samw size platters used in the old Prowler. I remember back in the late 90's at NAS Whidbey the Avionics techs would bring us crashed platters and we would media blast them removing the magnetic media (analog bleach bit)
were they then re-powdered?
@@highpath4776 nope thrown in the recycle bin and made into pop cans.
My god this is truly amazing. Watching the celebration when the disk worked was the cherry on top 😂
You got a sub from a couple hours south.
It is nice to see u so happy after all your efforts and hard work ....bravo
The sound of that disk spinning up reminded my of my old ICL Perq workstation that had a similar drive. You've shown it spin up, but how long does it take to spin down? I remember the floor shaking a few minutes after the Perq had been powered down.
This is fascinating, tech before i was born!! love it! keep on
Nice! The source files were an amazing find.
My bunny! After Usagi Yojimbo you are my favourite usagi of all time! I saw a 'games???' directory on there....I do hope that comes in for some early exploring once you have it restored again, as long as it dosen't ask you to play a nice game of chess....
Thanks!
That was actually one of the first files I opened up! I was really curious what was inside, but unfortunately the file is empty. There's nothing in there. Either it's on a bad sector or was something they were planning to write data to and never got to, but sadly it's just totally blank.
@@UsagiElectric Ah, what a shame. Glad you checked it before the drive died though, it would have driven me (and certainly you) nuts should the drive prove unfixable a second time!
"A strange game. The only winning move is not to play."
Great video, thank you so much for creating and sharing your knowledge.
This brought back a lot (A LOT) of memories diagnosing PC problems in the heydays of the 80s/90s...nothing has ever been quite like this era of technology.
I would often spend many hours researching or figuring out an issue, digging into the most esoteric of possibile causes... only to fix the problem, have it boot & run successfully & then develop a new problem about 30 minutes later 😅
I mean, it hasn't changed that much. Just the information's gotten a bit easier to access.
Software issues especially sometimes make me feel like I'm playing Whack-A-Mole.
"We're making progress. The error message has changed."🤣
I remember in 1978 when entry to our school computer club was achieved by buying an 8" disk. I still have mine somewhere!
To dump the drive, it would be interesting to run a read test (or make a "dd" like programme) but then actually tap all the signals of the drive into a logic analyzer/scope and share the waveform of the complete run. It would provide not only raw data, but checksums and low level geometry (as we can see the index signal).
Poor kitty had to be in the cone of shame. Hope he's well now
Always nice to end the Sunday with a Usagi Electric video! Meh, too bad the Finch 'died'. I'm still very impressed with your fault finding and the progress you already made. Keep going, we love it!
The following is from Chat GPT....
"Here are a few examples of capacities for different models within the CDC Wren series:
The CDC Wren I had a capacity ranging from 63.5MB to 340MB, depending on the specific model.
CDC Wren II: The Wren II, released in the late 1980s, offered capacities ranging from 130MB to 330MB. This was a significant leap in storage capacity compared to earlier hard drives.
CDC Wren III: The Wren III models, released around the same time, featured capacities that typically ranged from 540MB to 1.2GB."
My Wren CDC was 155mb. Hooked up to my Amiga 500 running a BBS @ 2400 baud.
8:25 I'm a big fan of the squeeze bulb desoldering iron too. It takes a lot of practice to be good with it though.
Ha... I, too, have a legacy 8" drive. It is a Hitachi 168MB. Weighs in at about 35#. It never ran. Cost me $3300 in 1981. It's current use is to weigh down out picnic umbrella.
i did enjoy seeing you get a win .it was bitter sweet but a good day :)
that bit when he waved his arms around after the heads self seeked I honestly thought he was going to punch it off the desk😂😂