Discussion in 'Retro & Arcade Worklogs' started by iMic, Oct 19, 2017.
Hopefully it's not a CS4235!
Be tempted to keep it and find a working case front for it. IBM stuff is getting collectable, particularly if you have screen/keyb/mouse etc.
Even those have a place in my heart!
But yes it's a nice machine, might be worth playing the long game and waiting for a case to appear.
Does Midori work on Win95?
I'm looking around for a replacement, but not finding anything. It'll be a matter of playing the waiting game and hoping something comes up.
(Who would have thought that finding parts assembling an IBM XT could be done in a year, but finding parts for a Pentium MMX era machine would be a several year process?)
Someone suggested looking into manufacturing a replacement panel. I may actually look into it. The machine is one part away from being a complete system, and I've seen more complex 3D printed structures made before, although some features like the textured panel would likely be removed.
How the machine looks without it...
EDIT: The onboard graphics is an ATI 3D RAGE II.
Also the Crystal sound chip is a CS4237B.
Can't hurt to try
I removed the door tracks and stuck them together with an ABS adhesive. It worked, but it cracked in some other places. It's turned extremely brittle. The outer cover, with the front handle, has cracked around a screw boss as well.
I'm not all that interested in keeping it at this stage, but I won't throw it out - someone will surely want a functional P166 MMX that doesn't care what it looks like, provided it works.
I suspect the older Aptiva 350-style bezel would be directly compatible as well, and possibly a stronger construction since it doesn't need the complex door mechanism:
But I don't suppose one of those will come up around here any time soon either.
Sold the Aptiva. Couldn't repair the front cover economically.
Next round of experimenting - IBM ThinkPad 370C and 755C. These machines look awesome, and are in decent condition too.
But - they won't run Windows 95. Because IBM screwed the BIOS. Something about an incomplete Video BIOS implementation. No matter what driver is loaded under Windows 95 for the Western Digital video controller, it's never able to run higher than 16 colours.
The Sound has a similar issue - it uses a Crystal chip, but with non-standard device IDs and some other BS that requires a custom IBM driver for Windows 3.x, that technically does work under 9x, but without support for MIDI and some other common functions.
IBM released a BIOS update for this machine to improve compatibility under Windows 95, but it won't flash without a fully charged battery - yep, it checks for one - and absolutely nobody manufactures replacements anymore. I partially disassembled one battery, but the casing is so fragile it would be extremely difficult to repack one without breaking it. Not to mention the high cost of the 10+ NiMH cells needed to rebuild it. Community driven patches were made to bypass the check on later model ThinkPads that had the same limitation and with newer versions of the flash utility, but nothing for the 370C / 755C.
Also, according to sources online, apparently the latest revision BIOS (v1.43) adds a startup check for an IBM-approved hard disk, so installing the latest BIOS update disables the use of third party drives and CompactFlash adapters. Once the stock drive dies, the machine dies with it.
There's one possible solution to the problem - with a nominal voltage of 9.6v required to drive the machine, it's possible I could add an 8x AA battery holder and toss in 8 Fujitsu 1.2v Ni-MH rechargeable cells, to see if that satisfies the battery check. Then flash BIOS v1.42, which should be new enough to support Windows 95 (support was added in v1.40) while still being old enough to support any hard disk.
It wouldn't completely resolve the sound limitations, but it may lift the restriction on the video hardware and colour palette a little, and make the machine usable.
But if there's any kind of intelligent battery circuitry in there (and face it, there probably is) that can't otherwise be transferred across from one of the dead packs, then it's all over. Another two machines get crossed off the list in the never-ending quest for a computer that can successfully run Windows 95. (Should still be able to make a machine that runs Windows 3.1 like a champ though.)
At least one tip has become apparent from all this - when it comes to Windows 95 compatible notebooks, stick with Toshiba or Compaq. At least their hardware is reasonably standard.
But ending on a brighter note, I may have a lead on some standard AT desktop components that I know will work, provided I can get ahold of them.
A PSA for anyone else out there:
It appears that Toshiba notebooks are becoming harder to find in decent condition with age, and not just due to general wear and tear.
I've purchased 4 of them so far, and all have cracked hinges or severe cracks around the outer casing. I assumed these were damaged, either over time or in transit, since most were dirty, had scuffs, etc. But yesterday I picked up a Tecra 700CT in immaculate condition, with bag, accessories and cables. From the outside it looked perfect.
I inspected inside the machine (and to remove the CMOS battery, which was leaking), and found the hinges broken in the same places - where the hinges screw into the brass inserts in the plastic top cover.
I've concluded that the cases are cracking without any kind of impact. The stresses of the brass insert (which is a press-fit) pressing on the incredibly thin (1mm) plastic, along with movement of the screen as it opens and closes, is enough to shatter the shell. This applies to most grey-case 486/Pentium era Toshiba notebooks, from experience this includes the Satellite and Tecra, and possibly the Portege and Libretto (although my Libretto seems to have held up so far).
I see these machines on eBay from time to time for several hundred dollars - being retro equipment and all - but I suspect that anyone paying these steep prices for a clean, good condition one would be disappointed given that this can happen to just about any machine, no matter how well it's been treated - or how well it was packed for shipping, for that matter.
I wanted to rescue this machine. It has two cosmetic cracks that both sealed reasonably well with some shockproof glue, but the structural crack around the brass plastic insert the hinge secures to was harder to tackle.
But a technique I learned around fifteen minutes prior ended up being the perfect solution - Baking (Bi-Carb) Soda and Super Glue.
I stuck the remaining pieces back around the insert. Then, I covered them in bicarb soda, pressing the powder into the gaps and cracks between the pieces. Then, a few drops of shockproof super glue over the top to set it in place. Within seconds it dried rock hard, and after leaving it for 15-20 minutes, I sanded off some of the excess and cleaned up around it.
That seems to have done it. The hinge moves fine, and shows no signs of weakening.
Next, out came the CMOS and memory backup batteries. Both were starting to leak, leaving some corrosion that needed to be cleaned up. Neither are needed for the machine to operate, and setting the date and time on startup is hardly an inconvenience.
Reassembling the machine was a pain. A couple of screws were positioned in such a way that required inserting the screwdriver into the head on an angle to clear hinges and brackets. That means the screwdriver skipped a few times; I'm aiming for "as good as possible" here, and damaged screws aren't welcome. Thankfully neither rounded out.
Super glue and I don't mix. Before I take the cap off, I set down some cloth, some plastic sheets, and a tarp across the floor. (Perhaps not that last one.) Somehow, it still ends up on my hands, my shirt, the walls, ceiling...
...and the lid of the computer. Because apparently some had transferred through the cloth onto the sponge I set down to avoid scratching and marking the lid. Because glue stains are much better than a few scratches, right? But even that thankfully came off with some JIF cream and Super Cheap Auto house brand glass cleaner, without so much as a trace it ever happened, so no harm done.
And as is customary around here, and I prepared to mark the completion of a difficult job with a "one finger salute", I realised that once all was said and done - the machine had gone back together, which is more than can be said for some other machines I've worked on. A little battle scarred and patched up, but assembled and working.
Pentium 120MHz, 16MB RAM, 4GB HDD. Active Matrix 800x600 TFT LCD. Currently awaiting an install of Windows 95.
Sweet little lappy! Had one like that for naval ops myself s long time ago. I remember having red alert on it.
Your whole post reminded me of the build experience itself which back then, without internet and computer literate friends was really a struggle in itself lol
I remember those rubbish pcima modem carts, and the rubbish fidily dongels that would come with them. The laptop itself was quite robust as was the floppy drive that came with it. (Mine came with win95 that you had to make a backup off on about 15 floppies) which I back then used on quite a few different systems lol.
Another Windows 95 machine. It's missing a few components at the moment.
It's a Digital Equipment Corp. Celebris 5100, marketed overseas as the Venturis 5100. Normally it looks something like this:
Configured with a Pentium 120MHz (incorrect for this system, as it appears to be an under clocked 200MHz MMX), 32MB RAM, 4GB Hard Disk and onboard S3 Vision 968 graphics.
The reason it's only a bare chassis at the moment is because it's been sanded to remove several smaller rust spots around the case, and is now currently at an electroplater receiving a zinc coating.
Hopefully the frame work should be completed sometime this week, then it'll be ready for reassembly.
And then, in addition to the above, this turned up.
An IBM Personal Computer 330. Pentium 133MHz, a massive 80MB RAM, 2.5GB HDD and onboard S3 Trio64V+ graphics.
The case is a little beaten up in some places, but the plastics are in decent shape, it's mostly rust free, and the few dents it has should tap out.
What a challenging machine the IBM PC 330 ended up being.
Starting with the front panel. Three of the four threaded screw posts that affix it to the steel lid had split. This time though, some baking soda in the cracks and around the post, followed by a few drops of super glue, and the posts are fixed stronger than new. The screws tightened down fine, and after tapping out a couple of dents, the front panel and lid are fixed.
The machine ran, but for some unknown reason, the Floppy Drive was permanently in write protected mode, even when testing with another drive and cable. Diskette write protection was disabled in the BIOS, but it ended up being a second write protect DIP switch on the motherboard that was set incorrectly.
Then the Floppy Drive worked when booting from an MS-DOS boot floppy, but not when booted into Windows 98 or when restarted into MS-DOS mode from within Windows 98. This was an interesting one. The drive worked, provided the Windows 98 CDROM wasn't in the drive during startup. The reason is because when the Windows 98 CD presents that "Boot from Hard Disk" and "Boot from CDROM" menu during boot, it creates a virtual drive in memory and mounts a "CDBOOT" image that provides the menu and basic DOS utilities. This image wasn't being cleared before handing over to Windows on the hard disk, and so the machine booted up with the remapped virtual drive in place.
The giveaway was when attempting to access Drive A from within My Computer, the floppy LED didn't illuminate - but the CDROM LED did. Simple solution - remove the Windows 98 CD from the drive and reboot.
After installing the newest BIOS update, chipset drivers, network drivers and a generic USB Mass Storage driver, the PC 330 is running Windows 98 smoothly.
Onto the DEC machine next, and maybe a couple of others depending on what happens this week.
The Digital Celebris frame is back from the electroplater.
The card selection is mostly locked in - Creative Vibra 16S OPL3 (ISA), 3Com EtherLink III (ISA) and Cirrus Logic GD-5446 (PCI). The CPU selection is up in the air, since I'm considering switching in an original Pentium with integrated heatsink and fan to simplify the case wiring.
Ordered some brand new screws from a distributor in Queensland, so once those come in, the reassembly can begin.
that's some commitment there - obligatory "how much did that cost???"
$30 for the electroplating, another $25 shipped for the screws. The plating was done through a regular contact of mine, and I did a lot of the sanding and prep work beforehand.
The screws arrived this afternoon, so I'm hoping to start reassembling it tonight.
Man, that electroplated chassis is magnificent! I love that you go the extra mile.
so worth it, that's stuff all really.
1:04 AM, and the machine is running stable.
DEC motherboard, Intel Pentium 90MHz processor (Socket 5), 512KB External Cache, 64MB RAM (2x32), 2.6GB Western Digital Caviar (SpinRite; All sectors OK), Creative Vibra 16 CT2800 (Yamaha OPL3), 3com EtherLink III ISA, Sony 1.44MB Floppy, Mitsumi Electric 8x CDROM.
The machine presented some stability issues after reassembly.
It sometimes stalled when booting from hard disk, and presented a "No bootable device found" message. Removing the 512KB cache module corrected this. When installing Windows, starting the installation took a long time, and it stalled before it started copying files. On the second install attempt, it completed successfully, but crashed often when performing operations from within the OS, such as playing back a video. When the original CPU was reinstalled, despite working before, it wouldn't complete the POST.
I tracked down some old Digital service manuals and technical service bulletins, and discovered the problem was five-fold.
The Celebris 5100 motherboard requires a cache module to operate reliably.
The BIOS automatically defaults to "No" for the Plug and Play Operating System setting when the CMOS is reset.
The BIOS doesn't automatically clear the hardware configuration when the CMOS is reset, it needs to be done manually from within the BIOS setup.
The Mitsumi CDROM laser needed cleaning.
Due to a combination of the above issues, the Windows installer copied corrupted system files during the initial setup.
After cleaning the cache module and motherboard slot, correcting the BIOS issues, cleaning the CDROM laser and reinstalling Windows, the computer started running reliably.
One card slot remains open at the moment, for an upgraded video card to be added later. The Cirrus Logic GD-5446 continues to be an option, but some other cards are being considered at the moment, including an ATI Mach64, an S3 Trio64V2/DX, S3 ViRGE/DX or a suitable 3DFX card, depending on compatibility, suitability for the build, and availability.