Ok first, I have to say, this video like all your videos is amazing. Your perseverance is inspiring! I can't believe how frustrating that machine is though. Serious credit where credit is due, I would have given up long ago!
As someone who worked in technical support at Intel when this system was on its very final days of official support - I can firmly say this is the worst 486 computer ever made. This system *IS* the Lovecraftian horror. I had one regular-caller customer who ran one of these as a NetWare 3 server as late as 1999, who constantly tried to get support on it. (He also had a newer system that WAS still supported, that he would open a ticket about, then when he got to level 3 support (me) he'd switch and start talking about this system.) Side (much longer) note: This system's existence helped pay off my college student loans. I worked for Intel over Y2k. Intel had already determined that all of the products with currently active support would handle Y2k just fine; but a few "recently support-expired" products *MIGHT* be impacted. This was one of them. The specific case was that internally, the system rolled the date over one digit at a time - so internally, 1999 would become 1990, then 1900, then 1000, then 2000. Only a few clock cycles of the CPU between each, but there was that tiny fraction of a second that the internal hardware clock would read an invalid date. Most operating systems no longer relied on the hardware clock for datekeeping while the system was on, and would sanity-check the results anyway. But.... Novell Netware 3 was one of the operating systems that actually made direct hardware calls to the clock regularly. And we knew there were customers who still ran it. So, in the extraordinarily unlikely event that an already obsolete system running an obsolete operating system happened to do a "date check" during those microseconds that the hardware would have an "invalid date" - there was a chance that the OS might crash. So, because Intel didn't want to appear to be "doing nothing" - they put level 3 technical support on high alert. For one week before Jan 1 to one week after, our level of support would *ONLY* accept "confirmed Y2k related tickets" from level 2 support. AND we would have at least one person on-premise sitting at a desk ready to accept these tickets 24/7 (normally level 3 support was M-F, 8-6.) *AND* all members of the team were on-call for the whole time. Required to carry a pager (remember those?) and be able to be in the office in one hour from being paged. I was a contract employee at the time, not a direct Intel employee. Only direct Intel employees could be in the building after hours unless accompanied by a direct intel employee. So in our department of 4 people, 2 of us couldn't be the only person there for 24/7 coverage. So the poor 2 direct employees had to split the overnight coverage that whole time. (Our boss joined them to at least add SOME extra leeway.) Oh, and they were all salaried. Us contract employees were hourly. Which is how paying off my student loans comes in to play. It turned out that the employment agency I worked through counted "on call" as the equivalent of being in the building working. So I clocked in Monday at 8 AM, then didn't clock out again until the *NEXT* Friday at 6 PM. After 8 hours, I was on time-and-a-half, after 12 hours, I was on double-time. And my boss had given me a 20% pay raise just a couple weeks before. This was split across two paychecks (I had to call and talk to my employment agency to confirm that I hadn't just forgotten to clock out and then took two weeks off.) But it paid off my student loans.
Oh man, this reminds me of the twice annual all hands of deck Ubuntu release process. Being salaried sucks when you have regular crunch time hours. The SystemPro is stranger (multiprocessor 486), but honestly, I was just waiting for the true horror to show up. I honestly am happy to get this thing off my desk for the time being, but I know it will return. I left a lot on the cutting room floor like the "i485 TurboCache module" and just focused on the worst parts of this system (the original script was like 12 pages long, I had to trim it considerably). Based off the rules, it sounds like you were out of Hillsborough as an orange badge?
Were there actually any legit y2k tickets come through? It was pre-internet for me, and I thought it (y2k bug) was bs from the get go. But stranger things have happened.
@@der0keks As someone who was there for the Y2K rollover, there was a lot of work done to be sure it wasn't an issue that day. If that work hadn't been pushed before hand it would have been very problematic. The regular media hyped it up after almost all of the work was done.
Oh man, the dos version of Norton Utilities. That takes me back. Norton used to be a great set of utilities, before it became the bloated behemoth mess it is these days. I used to sit and watch the defrag for fun. It was always oddly satisfying to watch.
I remember downloading the QNX "Incredible 1.44M Demo" back in the late '90s/early '00s. Thought it was pretty cool to have a floppy that booted up to a full GUI even if the usage was pretty limited.
I had that disk in my toolkit for years. It got lost at some point and I couldn't find the download any more. Is there an image available somewhere now?
QNX is rad, I remember that boot floppy demo. It was originally created by a Canadian company and powered these monstrosities called the Unisys ICON back in the mid-80s. Eventually QNX was bought out by another Canadian company RIM/Blackberry! (I think it got sold to someone in between there actually, 'cause it's used in lots of automotive applications)
Can you please drop me an email about this? (go to my Channel page -> About -> Contact Email). There's some things I want to ask about that I can't put in a public comment.
QNX was one of the operating systems I discovered back in 99/2000 when I went down the alternative OSes rabbit hole after discovering BeOS 5 personal edition. I got really excited for Blackberry 10 devices because Blackberry developed it on the QNX platform.
A couple years ago I removed an IBM PS/1 from daily database service since it was brand new. Imaged the drive, tossed it onto a new machine and a little bit of DOSBox later the ancient DB program still runs happily
These episodes/parts about this machine have been fascinating, thanks for the hard work. I hate to say it...when you asked what to do next with the machine I replied without thinking, “it’s time to throw it back!” I’m not sure how much more you can beat your head against this old thing. 😂
Your perseverance through this is definitely commendable! An absolutely chaotic machine. Would love to see you tackle compiling Linux for it, but I don't think anyone would blame you for just putting it aside on display as "Least compatible 486"!
You know, I do wonder if RUclips actually takes scheduled videos into account when the algorithm decides to promote things because both Parts 1 and 2 got a modest bump in views right before this went live.
@@NCommander This is actually quite an interesting thought, and it would make sense: If there's a Video RUclips plans to show to everyone, make them watch the first parts of the series too. Especially right before the new part releases, so the people are hyped for the new part. Good video by the way, just a bad computer. I'd love to see NetBSD boot on it.
This is brilliant, what an intense 3.11 install/"desktop" Regarding digital forensics: Have two HDD images of laptops used within an engineering firm, one was a Toshiba 200CDT that was connected to and used to program some PLC stuff with the KEPware program. Kepware is still around. The other was an AST Ascentia J50 that has a documentation program from Mitsubishi Electric called MELSEC MEDOC. I find it really interesting to see all the complex and advanced programs that was thrown onto the x86 (or equivalent) platform even in the early days of desktop computing. And that both ran on the somewhat unstable first release of Win95! I'll end this somewhat long comment by giving a thumbs up to all your interesting videos and topic, and also an honorary mention to that excellent HP monitor to show it all on - no capture card needed! ≈T
If you still have this system, I would be more than happy to rehome it. I have a couple machines from the 80s and early 90s that are workstation and server class. And always find their quirks fascinating. I can see it has the mezzanine connector for co-processor addon's, namely a Weitek floating point unit. Absolutely fascinating!
Once again an outstanding video. It's like you're doing in retrospect all the things I did in the 80's and 90's in real time. If you set up Unixware on a machine, things will officially be spooky
So, UnixWare *is* on my list, and was something I was going to do w/ this machine, but see the SCSI problems. I wanted to see the NetWare Approved UNIX and gag, and describe the long and crazy story between Ma Bell, Novell at SCO. I also done a lot of pre-RUclips Xenix work which will re-appear here, and I helped another retro youtuber on some recent SCO UNIX video projects.
I wonder if NetBSD and Linux are failing to boot because of a lack of RAM. From memory, both have compressed kernels which can be quite large when decompressed into RAM to run. It'll be interesting to see if adding more RAM makes a difference. Good luck!
Linux starts, it just doesn't see the SCSI controller. NetBSD is likely a probably in the real mode -> protected mode jump just based off what I can determine.
HI, by chance I come across your videos about LP486E; it was my first "real PC" in the 90's. I used DOS 6.22 first, than Win 3.11 but I manged to upgrade to Win 95OSR2 without problems; my LP486E worked with the "NCRSDMS" driver included in Windows 95 that had to be installed manually as not recognized automatically during setup, I remember well. Then, the other problem was only the LAN Driver, I got a driver written by a group of fan of this machine. Hope to be useful, thank for the videos that brought me 30 yrs back ...
Seing QNX again brings back some memories. I only tried QNX 4 with the demo disk, but I used QNX 6 through 6.2 as my main desktop operating system for over a year. Photon microGUI was very fast and lightweight and although the Voyager browser didn't render many sites corretcly, it was good enough for me. My computer back then was too old for Windows and even on Linux and OpenBSD, running Netscape was just too bloated for that old Pentium machine, even with the MMX and RAM upgrades. Fortunately I was able to buy a new computer in 2002 and Phoenix (now Firefox) came out a little after that.
it is a frustrating deep dive into the rabbit hole but somehow I'm looking forward to see it running a custom built linux and make the machine "usable" even today...
The NT4 setup script can be modified to allow a lower amount of memory to install. Change the value from 12288kb to a lower number, to allow the installer to install in less than 12mb.
I can use it for DOS and Windows 3.1 stuff which is probably what I'll end up doing with it. Nothing flashy but while LGR has the Woodgrain 486, I have the Pizzabox Basketcase!
You went through much more hell than I would have been willing to go through just to keep using a SCSI drive. I would install some solid state job like an IDE to CF or SD card adapter and used the IDE channel and disabling or bypassing the SCSI interface either through onboard jumpers or bios setting. I'm sure the solid state options would be much faster than using the SCSI over a normal IDE drive anyhow
As always... Thanks a lot for the profound insights. It was really a pleasure to listen to you... Especially when you started mentioning QNX :-D Nicely done!
I actually was inspired to do some digital forensics on old computers in my possession, not limited to the Compaq Presario desktop I got at a thrift store (and a hard drive taken out of a Dell Dimension 4100 that an old teacher of mine gave me).
If it was my machine, I think I'd be going down the alternative SCSI card route and installing NeXTSTEP. I used to have a NeXT Turbo Color machine, and it was wonderful for developing on.
Presumably most viewers of this ThemTube channel do have some sort of historical IT experience, but most of you forgot that in theses days the challenge was to get everything properly configured and setuped with incompatibilities, resource conflicts an proper driver support. 🧐
QNX is also used, for example, in the current generation of Sync 3 in-car entertainment systems by Ford. It succeeded the Windows CE-based Sync 1 and 2, and will be succeeded by Android in a couple of years.
What about gentoo? Would it run? I assume an older version starting from stage 2 or even stage 1 could be interesting. If there is enough energy and time also a LFS session could be possible.
At this point, I'd have to solve the underlying kernel issue before I could make Linux work (or just install an IDE drive and be done with it). Modern gentoo (if you're patent) can still be run on a 486, so I kinda want to go for something a bit more unorthodox.
Gentoo guy here. I tried that recently. Did not go well. The problem you'll find is modern tools are just WAY too heavy for an old box like this. You could pre-compile and run the binaries on older hardware, but there are a lot of features that are a given now, like SSE, functionality in later chipsets, etc. That makes it difficult for modern kernels to run at all, and some ecosystem software won't compile (or will compile, but won't run) without CPU features that weren't available yet due to inline ASM. So you would have to find a more era-appropriate snapshot, which is tough. The Gentoo archives don't go far enough back -- at least not that I found. And then what? The Portage tree is in version control, and I think you can go back a long ways, but it's just a catalog. You would need the source files for a lot of packages that might or might not still exist. I've had trouble finding "old versions" of software from 5 years ago, much less 199x. I think you would have better luck with a time capsule distro that has a CD release with the package repo available on it.
"Oh, Turbo Basic, I see... But the Spectrum was stupid fast compared to that piece of Architectural Archeology. I must forgot witch letter I used for the DVD before flashing the Bios... Still tryin' to get your voice over recognizable by the system manager? I don't know, Dude, it must be the DVD tray... QNX? Well, I liked so much those summer camp stupid check the box tests..."
When you start with the Pentium 2s and 3s and later, where they've been on broadband internet the whole time... that's when it will get interesting. A 486 was merely better at running Windows 3.1 and fairly basic versions of MS Office than a 386.
The only hope I can muster is to somehow build a recent version of Linux kernel and somehow enable OpenFirmware Device Tree support on x86 platform, then either let GRUB chain-load Das U-boot or let GRUB itself to feed the kernel a device tree that specifically tells each driver where to look for things. Basically I am treating this machine as a x86-based embedded system and using a bunch of techniques commonly found on ARM platforms.
It's a little more involved, but not much. I need to do an EISA ident lookup to see if the machine system board is INT3010, and then initialize the NCR 53C700 driver at the correct offset. The NCR driver is modularization so its admitedly straightforward(ish).
@@NCommander I recall having to use DOS and some "magic(tm)" driver (sys) to set things up for linux. From there load-lin would work, or a DOS EXE version of grub. But, ultimately, _avoid EISA._ (my DEC alpha is partially EISA. It's one of those oddball things with EISA, and PCI.)
@@NCommander in 1998 EISA was really dd /dev/kmem and then look for the right parts of the bios :-(. Since I had multiple pizza boxes with working scsi I have no memory if this one worked or not.
@@NCommander "It runs netbsd". They already claim that when you can download the bsd through a serial cable and then boot it and only run from memory, without any driver support. I discovered that when trying a HP9000 with a PA-risc cpu.
I think Debian or Arch might be able to run with such little RAM if you cut out any uneeded drivers and recompile with only basic programs. OOM-Killer might be upset though.
@@zenithseeker7 I've got more RAM coming when takes this more in the realm of possible, but NFS root eats 2 MiB by itself, and a kernel with lp486e.o is another 1-2 MiB. It's not *impossible*, but tight fit is putting it lightly.
Its a good chip, but in the days before opensource, someone had to write a driver - and someone had to test a driver, and in the case of linux, someone had to break a driver, fix a driver, break it again, fix it again... This, in the manual was a lie: "Video drivers for SCO UNIX should be obtained from the respective UNIX distributor." They were not. They were only available via Intel Dial up support. The file 'LPUNIXKIT' is an Intel file And "Boot Diskette supplied for R2. SDMS Ver 1.6 drivers" is in that kit.
As cool as SCSI is, screw it if you can use a CF card. Install Windows 95 OSR2/B with BootGUI=0 so that you have FAT32 on DOS and CIFS if you need it. If the builtin LAN chip doesn't work with 95 get an ISA network card that does. You have two slots, one for a Soundblaster, i would have used an AWE32 just because and because some folks from the Vogons forum have written initialization tools for old machines that don't have PnP. And the other for whatever you fancy unless you need it for a working network card. I would have used it to make it possible to change the CF card to better experiment with different OSes and to easily get a bunch of games on there. And with a CF card you can easily dd yourself a definitive state that you can always come back to if you want to experiment with old software under 95.
CF-IDE won't work with NT 3.1. 3.1 in general has a crapton of IDE issues, so it really only works with SCSI. I could use a NIC + XT-IDE bootROM in the socket which (at the cost of netbooting) would solve most of the geomtry issues, but it's still kinda a pain in the ass. That leaves one slot free for either a soundcard, or SCSI assuming the onboard one remains dead but on the whole its just a giant pain in the rear.
After reading a few other comments and some more digging, I think the machine must have been backed up with ARCserve, and then re-loaded after the HDD was swapped.
Well, that was interesting... not sure I have anything to add, other than noticing that after you did the low-level format, it appears that you neglected to do the necessary filesystem format. -- I hie from the MFM era myself, and well do I remember my ST225's annual rewhacks. (Until I discovered that putting a fan on the campstove-HD meant it never lost its marbles again.) Alas, I understand this "dance with them that brung you" thing all too well.. many a time I've argued with some cranky piece of hardware until it gave in and behaved, rather than digging a known-good replacement out of the parts box. It's the principle of the thing, ya know? PS. First vid where I've been able to see comments at ALL in about a week. (Affected Chrome and SeaMonkey but not Palemooon... affected PCLinuxOS/KDE but not WinXP... wtf.) Perhaps a good omen??
@@NCommander Ah, okay. "Through the magic of filmmaking..." Anyway, It musta been interesting, cuz here I sat peering at all the tiny print trying to figure it out along with you. :D
By the way, regarding the percentage readout in MemTest, I have a feeling it's supposed to be read as Pass 1 %, Pass 2 %, but due to the lack of screen real-estate it's just all glommed together as Pass1% and Pass2%.
Pass is shown lower right of the screen, and modern versions of memtest indeed count to 99->100 like you'd expect. Also, why end at 193 (which is the point where it actually rolls over).
It is totally random. Well, maybe deterministic, but not easily guessed. I have run Memtest on a lot of old computers. It never ends at 100%. Where it stops, nobody knows. You just keep waiting and eventually "Pass complete, no errors." Yay!
you might need to either recompile the linux kernel with the right driver and command line options, or if the driver is already in the kernel, find out what kernel command line options you need to insert into your grub or lilo config.
The driver exists, it just is incompatible with this system. I went through the source code very indepth, and used properly compiled custom kernels and the Slackware NCR 53C700 boot disk. This will need a kernel patch to resolve.
@@NCommander do you know what i mean by "kernel command line"? some drivers need to be configured at boot time this way, ie if you need to give them a different address than the one it would look at by default. i seem to remember needing to do this with an Adaptec SCSI card i had in a 486, or it would show exactly the kind of errors you saw, if it didn't have those kernel command line options set.
@@iceowl I'm fully aware of the kernel command line :). In this case, I specifically need additional code to do an EISA detection for INT3010 (the EISA ID of this machine), and then if found, initialize to the correct memory offset. That's a kernel patch w/ a new 53c700 wrapper driver (which is how most of the other NCR using devices are implemented)
Good video. I don't have this Pro Workstation, so don't have experience with it. However I do love the pizza box style, it reminds me a lot of the Commodore 128D & 128DCR (I don't have those, only C-64's). Have you asked around in vintage PC forums about these issues you mentioned? I searched VCF & found these using search "Intel Professional Workstation" >> www.vcfed.org/forum/search.php?searchid=12047597 . You might find more hits just using "Professional Workstation" idk. And try other vintage forums. Also, you were wondering what to do with this PC. Here are some of my blanket suggestions: 1. So it looks like this has a 33mhz 486. Does this PC have an all-encompassing crystal oscillator clock (OSC) on the main board? What if you replace it with a 40mhz oscillator? Or whatever frequency that will get it to 40mhz. That should help speed it up a little without burning it up too much. Just an idea. You may need to stick a passive heatsink (with or without fan) on the 486. I assume it does not have an Overdrive slot. What was that empty socket on the left of the 486? Is that for an FPU? Or is it to upgrade the CPU? 2. I also searched VCF for some blanket 486 O/S ideas & found one thread > www.vcfed.org/forum/archive/index.php/t-13163.html 3. What about PC GEOS? i had that back in the day (1990's) for my 8088's & 286's, i.e. Geoworks Ensemble 2.0. But afaik the most current versions are with Breadbox/bluewaysw > github.com/bluewaysw/ now now known as open-source FreeGEOS. I have not tried it (yet?). It also used to be NewDeal before Breadbox. You should be able to find the original files of Geoworks Ensemble & Breadbox Ensemble at winworldpc.com , but I'm not sure if that's legal or what. I still have my PC\GEOS 1.2 & 2.0 installation floppies. 4. If you stick with WfW 3.11, perhaps try Calmira? It is a UI shell that changes 3.x to look like 9x or XP. I have used Calmira XP before on a couple of my Win 3.1 machines in the past & it was awesome. www.calmira.de Look for Calmira XP 3.33 . There is also an LFN version, but I did not have luck with that. Calmira is inactive though, since 2008. There was also a Vista version I think. 5. I had another idea brewing, but now I forgot! And by the way, I used to work for Intel Corp. I was at the Hawthorn Farms campus in Hillsboro, Oregon. By the time I got there in the early mid 1990's, the 486's were pretty much phased out & the Pentium 60's had already started. It's been 15 years since I left & have not visited it since.
I ddi actually ask a few things on the IPW, but I sorta like following it more. 1. This system is 486 Overdrive compatible, and people have infact dropped Am586 133Mhz's in here, and Intel's own Socket 1 upgrades are fully compatible. I might actually get a fair kick because most Overdrive systems got stuck due to slow backplane, but with this system being EISA, it should be a significant performance boost. 2. I've got plenty of ideas here, its just a matter of time, making floppy disks, and suffering. 3. I actually did some w/ work w/ GEOS to get powered to OpenWatcom when it got openeed up source. Breadbox is a pretty interesting topic overall since that's what AOL for DOS ran on. 4. I'm actually pretty partial to Program Manager. If I do a desktop replacement, it would likely be Norton Desktop; I might Michael MJD has already said enough to Calmira that I won't be able to add much although I really haven't used it since I was an early adopter of Win 95, then OS/2, then Mac OS and finally Linux.
@@NCommander ... 1. Ah ok so it's an upgrade slot. I have one 486 (VTech Laser SX/3 25mhz PC), originally bought around 1993, where I have an old Evergreen 586 upgrade (AMD 5x86-133) in the socket, running at 100mhz (due to the 25mhz). I may replace the OSC to 33mhz & see if it's possible to get the thing running at 133 overall, even if it's only all ISA. The machine did not have external cache to begin with but once I figured it out & bought the proper cache chips (10 years ago), it made it faster (for a 486). I even put this old machine back to work again at the office for several years, using Windows 98se. 3. Yeah true, considering that Breadbox & AOL are somewhat related, having both their origins with Commodore machines, i.e. BerkeleySW (Geoworks) GEOS64 (of which I also have for my C-64) & AOL used to be QuantumLink (of which I was subscribed there for a bit, my first taste of the "Internet" before the Internet), respectively. 4. Alright, Norton it is. There was this big thick box of OS/2 installation, all alone in a drawer, that I discovered at Intel at my workplace that I was tempted to take, but did not. Haha, did not want to get caught or anything. Have never tried OS/2. Good luck with the IPW.
@@robwebnoid57631. I wasn't clear, that's an SX slot, so if I had an SX, I could put a FPU there. DX processor installed, so it's empty. There is a spot for a L2 Cache, but it's propertiary (that's the big black void on the mainboard, you can see it clearly in Part 1), it's call the Intel i485 TurboCache controller. 3. GEOS on the Commodore is on my topic list, it's part of why I picked up a Commodore 128 4. OS/2 is fun, is painful.
@@NCommander ... I see, all clear now. Is that turbocache hard to find? I would assume it would make things a little bit snappier. Unless someone has reverse-engineered it. I have a 486 laptop with proprietary memory slots & unfortunately won't be able to upgrade the onboard 4mb to 16mb max. Will look forward to your GEOS64/128 video.
if you could actually pull it out and take good pictures of the floppy disk eject button so I can get an idea of what I need to 3-D print, it would be amazing.
Came for the forensics, stayed for the lovecraftian horror that is this computer. Personally, I would have long since let it go. There are simply much better 486 based computers out there and that's including the bizarre embedded 486 computers.
Meh, you can’t blame the kids. It’s like getting all your learning from books. You may know, but you’ve never heard it pronounced. Beside, he’s just going for points now. I would have chucked that POS long before now! He’s got some chops just pushing it this far. I’m impressed!
Honestly, DAWS is how I heard it as a kid, and it wasn't until YT did anyone correct it. To be honest, I did actually try to check how I'm saying it, but I have trouble hearing the WS sound, and I'm told that I have an accent which doesn't help. (I'm from New York, but I lived in New England so my accent is generic Northeast).
In consumer space, the two years of support is pretty much par the course, but a lot of workstation and server-class boards I'm used to seeing decade+ support.
It's a 486DX so no PFU nueeded. It does have 8 MiB of FDO RAM installed, *but* QNX needs a minimium of 8 MB Extendeded, and this system only has 7600k-ish or I would have done the QNX demo in real hardware.
@@NCommander You're right i was distracted by the empty socket beside the CPU and when you showed a diagram of the board which was labeled "Performance Option" or so.... What is this socket about ? Regarding the RAM so you tell to upgrade to 16MB FDO Ram does not help ? BTW, please continue we need to see this prob. strangest 486 Workstation booting an OS (The weirder the better ;-)
@@yorkan213swd6 Its a slot for a 487, this system had an SX shipping configuration, and there's a SX/DX jumper on the main board. It just happens when you have a 486DX, no co-proc is needed. There's also a very large empty spot for a L2 cache module, but it's properitary so the odds of finding it are about nil. I *will* do more iwth this machine, it just needs to be on the side for my sanity ATM.
Ok first, I have to say, this video like all your videos is amazing. Your perseverance is inspiring! I can't believe how frustrating that machine is though. Serious credit where credit is due, I would have given up long ago!
Not going to lie, we got pretty close to Office Space territory on this one.
Lies. :-) Adrian would've brought out the current detector and tracked down bad 74-series logic chips on the motherboard.
As someone who worked in technical support at Intel when this system was on its very final days of official support - I can firmly say this is the worst 486 computer ever made. This system *IS* the Lovecraftian horror.
I had one regular-caller customer who ran one of these as a NetWare 3 server as late as 1999, who constantly tried to get support on it. (He also had a newer system that WAS still supported, that he would open a ticket about, then when he got to level 3 support (me) he'd switch and start talking about this system.)
Side (much longer) note: This system's existence helped pay off my college student loans. I worked for Intel over Y2k. Intel had already determined that all of the products with currently active support would handle Y2k just fine; but a few "recently support-expired" products *MIGHT* be impacted. This was one of them. The specific case was that internally, the system rolled the date over one digit at a time - so internally, 1999 would become 1990, then 1900, then 1000, then 2000. Only a few clock cycles of the CPU between each, but there was that tiny fraction of a second that the internal hardware clock would read an invalid date. Most operating systems no longer relied on the hardware clock for datekeeping while the system was on, and would sanity-check the results anyway. But.... Novell Netware 3 was one of the operating systems that actually made direct hardware calls to the clock regularly. And we knew there were customers who still ran it. So, in the extraordinarily unlikely event that an already obsolete system running an obsolete operating system happened to do a "date check" during those microseconds that the hardware would have an "invalid date" - there was a chance that the OS might crash.
So, because Intel didn't want to appear to be "doing nothing" - they put level 3 technical support on high alert. For one week before Jan 1 to one week after, our level of support would *ONLY* accept "confirmed Y2k related tickets" from level 2 support. AND we would have at least one person on-premise sitting at a desk ready to accept these tickets 24/7 (normally level 3 support was M-F, 8-6.) *AND* all members of the team were on-call for the whole time. Required to carry a pager (remember those?) and be able to be in the office in one hour from being paged.
I was a contract employee at the time, not a direct Intel employee. Only direct Intel employees could be in the building after hours unless accompanied by a direct intel employee. So in our department of 4 people, 2 of us couldn't be the only person there for 24/7 coverage. So the poor 2 direct employees had to split the overnight coverage that whole time. (Our boss joined them to at least add SOME extra leeway.)
Oh, and they were all salaried. Us contract employees were hourly. Which is how paying off my student loans comes in to play. It turned out that the employment agency I worked through counted "on call" as the equivalent of being in the building working.
So I clocked in Monday at 8 AM, then didn't clock out again until the *NEXT* Friday at 6 PM. After 8 hours, I was on time-and-a-half, after 12 hours, I was on double-time. And my boss had given me a 20% pay raise just a couple weeks before. This was split across two paychecks (I had to call and talk to my employment agency to confirm that I hadn't just forgotten to clock out and then took two weeks off.) But it paid off my student loans.
Oh man, this reminds me of the twice annual all hands of deck Ubuntu release process. Being salaried sucks when you have regular crunch time hours.
The SystemPro is stranger (multiprocessor 486), but honestly, I was just waiting for the true horror to show up. I honestly am happy to get this thing off my desk for the time being, but I know it will return.
I left a lot on the cutting room floor like the "i485 TurboCache module" and just focused on the worst parts of this system (the original script was like 12 pages long, I had to trim it considerably).
Based off the rules, it sounds like you were out of Hillsborough as an orange badge?
@@NCommander looking at the thumbs up im guessing it's a yes 😄
Were there actually any legit y2k tickets come through?
It was pre-internet for me, and I thought it (y2k bug) was bs from the get go. But stranger things have happened.
@@der0keks Not a single legit ticket made it through.
@@der0keks As someone who was there for the Y2K rollover, there was a lot of work done to be sure it wasn't an issue that day. If that work hadn't been pushed before hand it would have been very problematic. The regular media hyped it up after almost all of the work was done.
Oh man, the dos version of Norton Utilities. That takes me back. Norton used to be a great set of utilities, before it became the bloated behemoth mess it is these days.
I used to sit and watch the defrag for fun. It was always oddly satisfying to watch.
you've just accidentally reminded me to defrag my hard drives
thanks
I remember downloading the QNX "Incredible 1.44M Demo" back in the late '90s/early '00s. Thought it was pretty cool to have a floppy that booted up to a full GUI even if the usage was pretty limited.
I had that disk in my toolkit for years. It got lost at some point and I couldn't find the download any more.
Is there an image available somewhere now?
QNX is rad, I remember that boot floppy demo. It was originally created by a Canadian company and powered these monstrosities called the Unisys ICON back in the mid-80s. Eventually QNX was bought out by another Canadian company RIM/Blackberry! (I think it got sold to someone in between there actually, 'cause it's used in lots of automotive applications)
I would have liked to run it on real hardware, but it needs 8 MiB of extended memory, and this only has 7166k free after boot)
BBX & blackberry 10 was forked from qnx/neutrino, another story of "what it could have been"
I installed many of these for Reuters in the 90s. Should have a resource cd with all the drivers somewhere if you want them.
Can you please drop me an email about this? (go to my Channel page -> About -> Contact Email). There's some things I want to ask about that I can't put in a public comment.
QNX was one of the operating systems I discovered back in 99/2000 when I went down the alternative OSes rabbit hole after discovering BeOS 5 personal edition. I got really excited for Blackberry 10 devices because Blackberry developed it on the QNX platform.
A couple years ago I removed an IBM PS/1 from daily database service since it was brand new. Imaged the drive, tossed it onto a new machine and a little bit of DOSBox later the ancient DB program still runs happily
This channel should be renamed That Being Said.
These episodes/parts about this machine have been fascinating, thanks for the hard work. I hate to say it...when you asked what to do next with the machine I replied without thinking, “it’s time to throw it back!” I’m not sure how much more you can beat your head against this old thing. 😂
My head is one of the hardest substances known to mankind :)
ARCServe! We still use arcserve where I work. It's awful. It feels like it hasn't been updated since 1998.
Your perseverance through this is definitely commendable! An absolutely chaotic machine.
Would love to see you tackle compiling Linux for it, but I don't think anyone would blame you for just putting it aside on display as "Least compatible 486"!
What a gem of a series I’ve stumble upon. 😍 Subscribed!
Glad you enjoy it! It might be awhile before this machine shows back on stream though
I would love to see what it takes to make NetBSD work.
Thanks RUclips for showing me Part 1 on this series just in time for me to finish the earlier parts to watch this Video now
You know, I do wonder if RUclips actually takes scheduled videos into account when the algorithm decides to promote things because both Parts 1 and 2 got a modest bump in views right before this went live.
@@NCommander This is actually quite an interesting thought, and it would make sense: If there's a Video RUclips plans to show to everyone, make them watch the first parts of the series too. Especially right before the new part releases, so the people are hyped for the new part.
Good video by the way, just a bad computer. I'd love to see NetBSD boot on it.
Very helpful digital forensics video for beginners
This is brilliant, what an intense 3.11 install/"desktop"
Regarding digital forensics: Have two HDD images of laptops used within an engineering firm, one was a Toshiba 200CDT that was connected to and used to program some PLC stuff with the KEPware program. Kepware is still around. The other was an AST Ascentia J50 that has a documentation program from Mitsubishi Electric called MELSEC MEDOC. I find it really interesting to see all the complex and advanced programs that was thrown onto the x86 (or equivalent) platform even in the early days of desktop computing. And that both ran on the somewhat unstable first release of Win95!
I'll end this somewhat long comment by giving a thumbs up to all your interesting videos and topic, and also an honorary mention to that excellent HP monitor to show it all on - no capture card needed!
≈T
I don't actually own a capture card. If you see direct capture, that's a VM :)
If you still have this system, I would be more than happy to rehome it. I have a couple machines from the 80s and early 90s that are workstation and server class. And always find their quirks fascinating. I can see it has the mezzanine connector for co-processor addon's, namely a Weitek floating point unit. Absolutely fascinating!
Once again an outstanding video. It's like you're doing in retrospect all the things I did in the 80's and 90's in real time.
If you set up Unixware on a machine, things will officially be spooky
So, UnixWare *is* on my list, and was something I was going to do w/ this machine, but see the SCSI problems. I wanted to see the NetWare Approved UNIX and gag, and describe the long and crazy story between Ma Bell, Novell at SCO. I also done a lot of pre-RUclips Xenix work which will re-appear here, and I helped another retro youtuber on some recent SCO UNIX video projects.
I Loved that part about the machines past. Would love to see more people do similar
I wonder if NetBSD and Linux are failing to boot because of a lack of RAM. From memory, both have compressed kernels which can be quite large when decompressed into RAM to run. It'll be interesting to see if adding more RAM makes a difference.
Good luck!
Linux starts, it just doesn't see the SCSI controller. NetBSD is likely a probably in the real mode -> protected mode jump just based off what I can determine.
HI, by chance I come across your videos about LP486E; it was my first "real PC" in the 90's. I used DOS 6.22 first, than Win 3.11 but I manged to upgrade to Win 95OSR2 without problems; my LP486E worked with the "NCRSDMS" driver included in Windows 95 that had to be installed manually as not recognized automatically during setup, I remember well. Then, the other problem was only the LAN Driver, I got a driver written by a group of fan of this machine. Hope to be useful, thank for the videos that brought me 30 yrs back ...
This was a fun iceberg video without the iceberg.
At least I only had to bail out to my desk chair.
Seing QNX again brings back some memories. I only tried QNX 4 with the demo disk, but I used QNX 6 through 6.2 as my main desktop operating system for over a year. Photon microGUI was very fast and lightweight and although the Voyager browser didn't render many sites corretcly, it was good enough for me. My computer back then was too old for Windows and even on Linux and OpenBSD, running Netscape was just too bloated for that old Pentium machine, even with the MMX and RAM upgrades. Fortunately I was able to buy a new computer in 2002 and Phoenix (now Firefox) came out a little after that.
it is a frustrating deep dive into the rabbit hole but somehow I'm looking forward to see it running a custom built linux and make the machine "usable" even today...
Just in time! Can't wait to watch this!
The NT4 setup script can be modified to allow a lower amount of memory to install. Change the value from 12288kb to a lower number, to allow the installer to install in less than 12mb.
I'm hear for the stories of Lovecraftian horrors
Subscribing this comment for the lovecraftian horrors
Lovecraft approved comment right here.
this is a super weird computer for sure!
I am in love with it's looks. If all fails this should be turned into a casemod.
I can use it for DOS and Windows 3.1 stuff which is probably what I'll end up doing with it. Nothing flashy but while LGR has the Woodgrain 486, I have the Pizzabox Basketcase!
You went through much more hell than I would have been willing to go through just to keep using a SCSI drive.
I would install some solid state job like an IDE to CF or SD card adapter and used the IDE channel and disabling or bypassing the SCSI interface either through onboard jumpers or bios setting.
I'm sure the solid state options would be much faster than using the SCSI over a normal IDE drive anyhow
As always... Thanks a lot for the profound insights. It was really a pleasure to listen to you... Especially when you started mentioning QNX :-D
Nicely done!
Thank you kindly :)
I'd love to see all of those suggestions, especially compiling Linux.
A very interesting approach to retro computing :)
I actually was inspired to do some digital forensics on old computers in my possession, not limited to the Compaq Presario desktop I got at a thrift store (and a hard drive taken out of a Dell Dimension 4100 that an old teacher of mine gave me).
If it was my machine, I think I'd be going down the alternative SCSI card route and installing NeXTSTEP. I used to have a NeXT Turbo Color machine, and it was wonderful for developing on.
It would be both replacement NIC and replacement SCSI card for NeXTstep, and giving up the sound card :/
@@NCommander I know it's a year later, but if it were me I'd just chuck the SCSI, and use the IDE instead. Why insist on more headaches?
Presumably most viewers of this ThemTube channel do have some sort of historical IT experience, but most of you forgot that in theses days the challenge was to get everything properly configured and setuped with incompatibilities, resource conflicts an proper driver support. 🧐
Loved the video. For an old SCSI like that I'd try using SpinRite from GRC to fix it. If the bios can see it SpinRite can.
The drive is fine, its just the actual SCSI controller on this system is hot garbage.
QNX is also used, for example, in the current generation of Sync 3 in-car entertainment systems by Ford. It succeeded the Windows CE-based Sync 1 and 2, and will be succeeded by Android in a couple of years.
What about gentoo? Would it run? I assume an older version starting from stage 2 or even stage 1 could be interesting. If there is enough energy and time also a LFS session could be possible.
At this point, I'd have to solve the underlying kernel issue before I could make Linux work (or just install an IDE drive and be done with it). Modern gentoo (if you're patent) can still be run on a 486, so I kinda want to go for something a bit more unorthodox.
Thinking of unorthodox i would suggest LFS (linux from scratch) or an old debian/slackware/redhat
Gentoo guy here. I tried that recently. Did not go well.
The problem you'll find is modern tools are just WAY too heavy for an old box like this. You could pre-compile and run the binaries on older hardware, but there are a lot of features that are a given now, like SSE, functionality in later chipsets, etc. That makes it difficult for modern kernels to run at all, and some ecosystem software won't compile (or will compile, but won't run) without CPU features that weren't available yet due to inline ASM.
So you would have to find a more era-appropriate snapshot, which is tough. The Gentoo archives don't go far enough back -- at least not that I found. And then what? The Portage tree is in version control, and I think you can go back a long ways, but it's just a catalog. You would need the source files for a lot of packages that might or might not still exist. I've had trouble finding "old versions" of software from 5 years ago, much less 199x.
I think you would have better luck with a time capsule distro that has a CD release with the package repo available on it.
slackware ZIPslack distro is the way to go, just boot from zipdisk and SLIP or PLIP for IP
Hmmm did you try disabling the L1 cache again? Seems the problems started after that ?
"Oh, Turbo Basic, I see... But the Spectrum was stupid fast compared to that piece of Architectural Archeology. I must forgot witch letter I used for the DVD before flashing the Bios... Still tryin' to get your voice over recognizable by the system manager? I don't know, Dude, it must be the DVD tray... QNX? Well, I liked so much those summer camp stupid check the box tests..."
When you start with the Pentium 2s and 3s and later, where they've been on broadband internet the whole time... that's when it will get interesting. A 486 was merely better at running Windows 3.1 and fairly basic versions of MS Office than a 386.
The only hope I can muster is to somehow build a recent version of Linux kernel and somehow enable OpenFirmware Device Tree support on x86 platform, then either let GRUB chain-load Das U-boot or let GRUB itself to feed the kernel a device tree that specifically tells each driver where to look for things. Basically I am treating this machine as a x86-based embedded system and using a bunch of techniques commonly found on ARM platforms.
I’d love to see Linux get patched to run this. Sounds like it’s just moving some memory addresses around
It's a little more involved, but not much. I need to do an EISA ident lookup to see if the machine system board is INT3010, and then initialize the NCR 53C700 driver at the correct offset. The NCR driver is modularization so its admitedly straightforward(ish).
@@NCommander I recall having to use DOS and some "magic(tm)" driver (sys) to set things up for linux. From there load-lin would work, or a DOS EXE version of grub. But, ultimately, _avoid EISA._ (my DEC alpha is partially EISA. It's one of those oddball things with EISA, and PCI.)
@@NCommander in 1998 EISA was really dd /dev/kmem and then look for the right parts of the bios :-(. Since I had multiple pizza boxes with working scsi I have no memory if this one worked or not.
@@NCommander "It runs netbsd". They already claim that when you can download the bsd through a serial cable and then boot it and only run from memory, without any driver support. I discovered that when trying a HP9000 with a PA-risc cpu.
Could you drop an Adaptec SCSI board in there. Adaptec boards are bomb proof with many different OS supported drivers
I second this. I had an EISA machine and bought a 32 bit SCSI adapter, since SCSI was a much better option than IDE drives those days.
If you would try compile a custom distro, remember to disable flags, such old hardware may crash with new gcc optimizations due to lower RAM etc...
Won't be custom, I'd take off the shelf Ubuntu (or similar) and bend it to my will :).
I think Debian or Arch might be able to run with such little RAM if you cut out any uneeded drivers and recompile with only basic programs. OOM-Killer might be upset though.
@@zenithseeker7 I've got more RAM coming when takes this more in the realm of possible, but NFS root eats 2 MiB by itself, and a kernel with lp486e.o is another 1-2 MiB. It's not *impossible*, but tight fit is putting it lightly.
That box begs for Gentoo.
@@NCommander What does it max out at? Given it's got the 8mb already, I'd expect at 16mb, but I've seen a few of the era that could take more.
A very interesting series. Nine months later without an update and I imagine you've run out of ideas to get the damn thing working properly.
It works as well as it did from the factory. I have issues putting more effort in getting it working than Intel did ...
so just out of curiosity what distro of Linux do you use because I use Arch Linux I run KDE as a desktop environment and I use the linux-zen kernel?
I'd recommend buildroot for making a Linux based "disto" to install on this machine.
Recompiling Ubuntu or another binary distro is the point though; going w/ Gentoo or buildroot isn't very challenging.
@@NCommander Ahh so you're in it for the challenge. Good man!
Its a good chip, but in the days before opensource, someone had to write a driver - and someone had to test a driver,
and in the case of linux, someone had to break a driver, fix a driver, break it again, fix it again...
This, in the manual was a lie: "Video drivers for SCO UNIX should be obtained from the respective UNIX distributor."
They were not. They were only available via Intel Dial up support. The file 'LPUNIXKIT' is an Intel file
And "Boot Diskette supplied for R2. SDMS Ver 1.6 drivers" is in that kit.
Did OpenBSD have support for that SCSI card? A lot of the BSDs share drivers.
TBH, I didn't check. Generally I was like "yeah, I'm out" when NetBSD tapped out cause its NetBSD.
I like the Ubuntu idea
Memtest answer, more tests more likely to be good....back when I was a technician I had a few boxes that were ok until 40 or even 50+ passes
oh this is gonna be hell, i just know it
The second half of the video is for you.
@@NCommander update: i was right
He pronounces DOS like a serial killer would
"On this episode of Forensic Files, learn how a vintage tech collector committed crimes against the English language"
As cool as SCSI is, screw it if you can use a CF card.
Install Windows 95 OSR2/B with BootGUI=0 so that you have FAT32 on DOS and CIFS if you need it. If the builtin LAN chip doesn't work with 95 get an ISA network card that does.
You have two slots, one for a Soundblaster, i would have used an AWE32 just because and because some folks from the Vogons forum have written initialization tools for old machines that don't have PnP. And the other for whatever you fancy unless you need it for a working network card. I would have used it to make it possible to change the CF card to better experiment with different OSes and to easily get a bunch of games on there.
And with a CF card you can easily dd yourself a definitive state that you can always come back to if you want to experiment with old software under 95.
CF-IDE won't work with NT 3.1. 3.1 in general has a crapton of IDE issues, so it really only works with SCSI. I could use a NIC + XT-IDE bootROM in the socket which (at the cost of netbooting) would solve most of the geomtry issues, but it's still kinda a pain in the ass. That leaves one slot free for either a soundcard, or SCSI assuming the onboard one remains dead but on the whole its just a giant pain in the rear.
Try installing Banyan Vines.
where can we download an image of the harddisk?
Incredible. So many years in use without proper reinstalling of the OS... A very weird machine.
After reading a few other comments and some more digging, I think the machine must have been backed up with ARCserve, and then re-loaded after the HDD was swapped.
Well, that was interesting... not sure I have anything to add, other than noticing that after you did the low-level format, it appears that you neglected to do the necessary filesystem format. -- I hie from the MFM era myself, and well do I remember my ST225's annual rewhacks. (Until I discovered that putting a fan on the campstove-HD meant it never lost its marbles again.)
Alas, I understand this "dance with them that brung you" thing all too well.. many a time I've argued with some cranky piece of hardware until it gave in and behaved, rather than digging a known-good replacement out of the parts box. It's the principle of the thing, ya know?
PS. First vid where I've been able to see comments at ALL in about a week. (Affected Chrome and SeaMonkey but not Palemooon... affected PCLinuxOS/KDE but not WinXP... wtf.) Perhaps a good omen??
I did do a regular format, it's just on the cutting room floor. A lot of footage is on the cutting room floor >.
@@NCommander Ah, okay. "Through the magic of filmmaking..." Anyway, It musta been interesting, cuz here I sat peering at all the tiny print trying to figure it out along with you. :D
By the way, regarding the percentage readout in MemTest, I have a feeling it's supposed to be read as Pass 1 %, Pass 2 %, but due to the lack of screen real-estate it's just all glommed together as Pass1% and Pass2%.
Pass is shown lower right of the screen, and modern versions of memtest indeed count to 99->100 like you'd expect. Also, why end at 193 (which is the point where it actually rolls over).
It is totally random. Well, maybe deterministic, but not easily guessed. I have run Memtest on a lot of old computers. It never ends at 100%. Where it stops, nobody knows. You just keep waiting and eventually "Pass complete, no errors." Yay!
So wait you're saying it's Windows NOT for workgroups?
Holy crap. That processor was made a week before I was born.
you might need to either recompile the linux kernel with the right driver and command line options, or if the driver is already in the kernel, find out what kernel command line options you need to insert into your grub or lilo config.
The driver exists, it just is incompatible with this system. I went through the source code very indepth, and used properly compiled custom kernels and the Slackware NCR 53C700 boot disk.
This will need a kernel patch to resolve.
@@NCommander do you know what i mean by "kernel command line"? some drivers need to be configured at boot time this way, ie if you need to give them a different address than the one it would look at by default. i seem to remember needing to do this with an Adaptec SCSI card i had in a 486, or it would show exactly the kind of errors you saw, if it didn't have those kernel command line options set.
@@iceowl I'm fully aware of the kernel command line :). In this case, I specifically need additional code to do an EISA detection for INT3010 (the EISA ID of this machine), and then if found, initialize to the correct memory offset. That's a kernel patch w/ a new 53c700 wrapper driver (which is how most of the other NCR using devices are implemented)
Good video. I don't have this Pro Workstation, so don't have experience with it. However I do love the pizza box style, it reminds me a lot of the Commodore 128D & 128DCR (I don't have those, only C-64's). Have you asked around in vintage PC forums about these issues you mentioned? I searched VCF & found these using search "Intel Professional Workstation" >> www.vcfed.org/forum/search.php?searchid=12047597 . You might find more hits just using "Professional Workstation" idk. And try other vintage forums.
Also, you were wondering what to do with this PC. Here are some of my blanket suggestions:
1. So it looks like this has a 33mhz 486. Does this PC have an all-encompassing crystal oscillator clock (OSC) on the main board? What if you replace it with a 40mhz oscillator? Or whatever frequency that will get it to 40mhz. That should help speed it up a little without burning it up too much. Just an idea. You may need to stick a passive heatsink (with or without fan) on the 486. I assume it does not have an Overdrive slot. What was that empty socket on the left of the 486? Is that for an FPU? Or is it to upgrade the CPU?
2. I also searched VCF for some blanket 486 O/S ideas & found one thread > www.vcfed.org/forum/archive/index.php/t-13163.html
3. What about PC GEOS? i had that back in the day (1990's) for my 8088's & 286's, i.e. Geoworks Ensemble 2.0. But afaik the most current versions are with Breadbox/bluewaysw > github.com/bluewaysw/ now now known as open-source FreeGEOS. I have not tried it (yet?). It also used to be NewDeal before Breadbox. You should be able to find the original files of Geoworks Ensemble & Breadbox Ensemble at winworldpc.com , but I'm not sure if that's legal or what. I still have my PC\GEOS 1.2 & 2.0 installation floppies.
4. If you stick with WfW 3.11, perhaps try Calmira? It is a UI shell that changes 3.x to look like 9x or XP. I have used Calmira XP before on a couple of my Win 3.1 machines in the past & it was awesome. www.calmira.de Look for Calmira XP 3.33 . There is also an LFN version, but I did not have luck with that. Calmira is inactive though, since 2008. There was also a Vista version I think.
5. I had another idea brewing, but now I forgot!
And by the way, I used to work for Intel Corp. I was at the Hawthorn Farms campus in Hillsboro, Oregon. By the time I got there in the early mid 1990's, the 486's were pretty much phased out & the Pentium 60's had already started. It's been 15 years since I left & have not visited it since.
I ddi actually ask a few things on the IPW, but I sorta like following it more.
1. This system is 486 Overdrive compatible, and people have infact dropped Am586 133Mhz's in here, and Intel's own Socket 1 upgrades are fully compatible. I might actually get a fair kick because most Overdrive systems got stuck due to slow backplane, but with this system being EISA, it should be a significant performance boost.
2. I've got plenty of ideas here, its just a matter of time, making floppy disks, and suffering.
3. I actually did some w/ work w/ GEOS to get powered to OpenWatcom when it got openeed up source. Breadbox is a pretty interesting topic overall since that's what AOL for DOS ran on.
4. I'm actually pretty partial to Program Manager. If I do a desktop replacement, it would likely be Norton Desktop; I might Michael MJD has already said enough to Calmira that I won't be able to add much although I really haven't used it since I was an early adopter of Win 95, then OS/2, then Mac OS and finally Linux.
@@NCommander ...
1. Ah ok so it's an upgrade slot. I have one 486 (VTech Laser SX/3 25mhz PC), originally bought around 1993, where I have an old Evergreen 586 upgrade (AMD 5x86-133) in the socket, running at 100mhz (due to the 25mhz). I may replace the OSC to 33mhz & see if it's possible to get the thing running at 133 overall, even if it's only all ISA. The machine did not have external cache to begin with but once I figured it out & bought the proper cache chips (10 years ago), it made it faster (for a 486). I even put this old machine back to work again at the office for several years, using Windows 98se.
3. Yeah true, considering that Breadbox & AOL are somewhat related, having both their origins with Commodore machines, i.e. BerkeleySW (Geoworks) GEOS64 (of which I also have for my C-64) & AOL used to be QuantumLink (of which I was subscribed there for a bit, my first taste of the "Internet" before the Internet), respectively.
4. Alright, Norton it is. There was this big thick box of OS/2 installation, all alone in a drawer, that I discovered at Intel at my workplace that I was tempted to take, but did not. Haha, did not want to get caught or anything. Have never tried OS/2.
Good luck with the IPW.
@@robwebnoid57631. I wasn't clear, that's an SX slot, so if I had an SX, I could put a FPU there. DX processor installed, so it's empty.
There is a spot for a L2 Cache, but it's propertiary (that's the big black void on the mainboard, you can see it clearly in Part 1), it's call the Intel i485 TurboCache controller.
3. GEOS on the Commodore is on my topic list, it's part of why I picked up a Commodore 128
4. OS/2 is fun, is painful.
@@NCommander ... I see, all clear now. Is that turbocache hard to find? I would assume it would make things a little bit snappier. Unless someone has reverse-engineered it. I have a 486 laptop with proprietary memory slots & unfortunately won't be able to upgrade the onboard 4mb to 16mb max. Will look forward to your GEOS64/128 video.
QNX! Wow!
I've got one - no idea where I got it - of these systems stored away too...
if you could actually pull it out and take good pictures of the floppy disk eject button so I can get an idea of what I need to 3-D print, it would be amazing.
Came for the forensics, stayed for the lovecraftian horror that is this computer. Personally, I would have long since let it go. There are simply much better 486 based computers out there and that's including the bizarre embedded 486 computers.
Lovecraft said the horror was under the ocean, not in my closet >.>;
@@NCommander Lovecraft never saw this computer then...
@@weirdmindofesh Its Lovecraft for the digital era :)
I think it's cool just as a QNX machine.
Every time I read "Digital Forensics" I think of Evidence Elininator... lulz When he gets out of jail he needs to make it for Windows 10.
Wait Norton was actually useful! And I grew up in the 1995 world of dial-up and windows 95.
200% should mean it went over it twice. God I hope that's what it means.
If I'm testing memory, minimum of 1000% or it's not complete.
Wow!
I'd bet this thing was in use less than 5 years ago...
intel is the The Computer Inside.
So this was used for a university for a few years then scavenged out of the bin and continued for a bit as a cheap system.
No SpinRite?
Spinrite doesn't help on SCSI drives, as stated in the documentation.
Someone find a way of porting the old 3.1 driver for him !
Assuming it actually works. I still have no way of testing.
Hey Is there a Part 4 ? :)
i never end up getting the cookie
No one knows exists...that's mean to all those people that had BB10 devices...oh wait...nope that is no one
“DOSS” not “DAWS”.
OK good it's not just me.. Love this channel but it causes physical pain everytime I hear "dozz" maybe a form of retro term synesthesia.. 😂
Meh, you can’t blame the kids. It’s like getting all your learning from books. You may know, but you’ve never heard it pronounced.
Beside, he’s just going for points now. I would have chucked that POS long before now! He’s got some chops just pushing it this far. I’m impressed!
Honestly, DAWS is how I heard it as a kid, and it wasn't until YT did anyone correct it. To be honest, I did actually try to check how I'm saying it, but I have trouble hearing the WS sound, and I'm told that I have an accent which doesn't help. (I'm from New York, but I lived in New England so my accent is generic Northeast).
@@NCommander Interesting, this might be poll worthy. Is DOSS vs DAWS a west vs east coast thing? Here in CA I always heard DOSS as a kid.
@@NCommander well I'm just giving you a hard time, I honestly don't care how you say it as long as you keep the vids coming! 😁
Atrocious support by Intel. Hmmm wait.. aren't we supposed to buy new motherboards every 2 years? Looks like a rather old scheme. ;)
In consumer space, the two years of support is pretty much par the course, but a lot of workstation and server-class boards I'm used to seeing decade+ support.
Boat anchor
Send the unit to e-waste... Problem solved...
It pronounced "DOSS" not "DAWS".
Go Bold + 8MB Ram & 487 FPU then QNX...
It's a 486DX so no PFU nueeded. It does have 8 MiB of FDO RAM installed, *but* QNX needs a minimium of 8 MB Extendeded, and this system only has 7600k-ish or I would have done the QNX demo in real hardware.
@@NCommander You're right i was distracted by the empty socket beside the CPU and when you showed a diagram of the board which was labeled "Performance Option" or so.... What is this socket about ? Regarding the RAM so you tell to upgrade to 16MB FDO Ram does not help ? BTW, please continue we need to see this prob. strangest 486 Workstation booting an OS (The weirder the better ;-)
@@yorkan213swd6 Its a slot for a 487, this system had an SX shipping configuration, and there's a SX/DX jumper on the main board. It just happens when you have a 486DX, no co-proc is needed. There's also a very large empty spot for a L2 cache module, but it's properitary so the odds of finding it are about nil. I *will* do more iwth this machine, it just needs to be on the side for my sanity ATM.
Pleassssssseeeee pleassssseeee stop saying "dAWs".... It is pronounced "doss".... Like "floss" but with a d.
I'd love to see all of those suggestions, especially compiling Linux.