That is a really nice little machine. I never realized they had ISA slots: I thought these were MCA. 😀

PeterLI wrote:

That is a really nice little machine. I never realized they had ISA slots: I thought these were MCA. 😀

Yeah apparently any PS/2 below '50' is autonomously ISA.

Of course it's a bit odd that IBM made it ISA (16-bit ISA no less) so late in the game. Probably due to the fact MCA was already 'out'.

I'm not complaining though, means I don't have to hunt down another MCA soundcard.

EDIT:

Hmmm... I just had an idea! I could make this my 'adlib gold' computer. Yeah, that's what I'll probably do.

Adlib gold would be a great sound solution for this PS/2 making it even more special! I like the way it's made, very much like IBM, overengineered to hell 😀

486SLC can't address more than 16MB of memory because of 16-bit bus, right? So that's the answer why this machine is ISA. 486DLC could have a chance of having something more advanced due to 32-bit FSB, but not SLC... The way I see it, it was a neat little machine for home and office use, nothing heavy enough to buy a true 486 (or Pentium-60!) for. With lowest power consumption possible and zero noise level it's very parctical, if you ask me. Is the speaker any good like in later IBM computers or is it standard?

Never seen that type of IBM PS/2 before, very neat stuff.

There is one tiny bit of MCA in there, the ISA riser card edge uses a surplus 16bit MCA slot (with video option). 😜

NJRoadfan wrote:

There is one tiny bit of MCA in there, the ISA riser card edge uses a surplus 16bit MCA slot (with video option). 😜

That would explain why it's brown and made by burndy.

So I finally got the Hakko, desoldering commenced on both the adlib gold and the model 33:

Removing the PS/2E's motherboard revealed buttloads of surface mount components on the bottom, that would be horrible to manufacture:

Here we have the board recapped with rubycons, chemicons, and sanyo oscons. Please note that upon removing two 22uFs and one 68uF, I did not note what was what, so I just made all three 68uF:

I wanted some nice beefy 470uF decouplers on the adlib gold, as they were slightly taller than the I/O panel, I made sure the PS/2E had sufficient clearance with the top of its lid, and it does:

And here we have it re-assembled (I had to disassemble it a second time as I accidentally put 220uFs in place of the 68uFs which were 10 volts...). It's humorous how the adlib gold overshadows the computer:

After the recapping surgery (sans the stupid surface mount tantalums at the bottom) the machine powers on and works without a problem. After putting the 1.4 GB HDD in with windows 95 ready to install, it looks like the computer can't see the HDD. I am not sure if this is because the VLSI chip now has 68uF chips (unlikely) if the IC died, if the DDO on the HDD broke (a little more likely), or what. I'll be doing some more tests later. Too bad... I was really looking forward to hearing the adlib gold after its massive recapping.

The computer has 12 MB of RAM (two 4MB SIMMs and 4MB onboard). I opted out on putting two 8MB SIMMs in just because I want to save as much power as possible... every chip counts when you only have 24 watts. If for some reason the HDD controller is pooched, that'll suck. I'll have to use an external HDD then.

Okay I have an update-- the problem is that the Adlib Gold (most likely one of its bus interfaces) is conflicting with the HDD controller-- as the IDE is linked to the ISA bus (after reviewing it with my friend). For a test in proof the HDD is booted into without problem once the adlib gold is removed (rather convenient removing it, I just pop out the riser with the card).

UNFORTUNATELY no one on the entire internet documented the adlib gold's jumpers, so I'll just play around with them until the conflict is resolved. Ugh...

EDIT:

Got the whole thing working nicely with 95 and the adlib gold. Will put up some recordings later. Probably on the weekend.

raymangold wrote:

UNFORTUNATELY no one on the entire internet documented the adlib gold's jumpers...

I do plan to scan the manual eventually, but for future reference, here's a transcription at least:

Ad Lib Gold User's Guide wrote:

To make it easier to configure the Ad Lib Gold card, we have made the Interrupt lines (IRQ) and DMA channels software selectabl […]
Show full quote

To make it easier to configure the Ad Lib Gold card, we have made the Interrupt lines
(IRQ) and DMA channels software selectable, thereby keeping the amount of jumpers
to a minimum. The four remaining jumper sets are the game port enable jumper, the
dual joystick selector jumpers, the port address jumper and the Control chip reset
jumper.

The Gold card jumpers are the following:

Game port enable jumper (JP1)
This jumper lets the user enable/disable Ad Lib Gold's on-board game port
interface. The interface should be disabled in cases where the user already has a
standard PC game port interface inside his/her PC, to avoid conflicts.

Jumper setting is shown here:

[-xx] Game port enabled

[xx-] Game port disabled

Dual joystick selector jumpers (JP2-7)
These jumpers let the user change the factory-set "joystick plus MIDI" option (all
jumpers on the bracket side, pins No. 2-3) to the "two joysticks without MIDI"
option (all jumpers on the opposite side, pins No. 1-2). All jumpers in this
selector must be changed to the same position, as shown in the following
illustrations:

[-xx] Single joystick with MIDI option (factory-set)
[-xx]
[-xx]
[-xx]
[-xx]
[-xx]

[xx-] Dual joystick option
[xx-]
[xx-]
[xx-]
[xx-]
[xx-]

Port address jumper (JP8)
This jumper lets the user choose a single or double port address for the Gold card.
The Ad Lib Gold addresses can be assigned by software programming. The
default port address of the Gold card is 388H and can be changed by software in
cases where another card inside the PC uses the same address, in order to avoid
conflicts. In the case where the software cannot recognize the programmed
address, the port address jumper is used to force the Gold card into answering at
both the programmed address and the default factory address 388H.

NOTE: The port address jumper is factory set to a single port
address position (jumper plugged on the two upper pins) which
enables only one port address to be used at a time.*

* It should be kept in the single port address position. Double port address position should only be
used in maintenance procedures.

Control chip reset jumper (JP9)
This jumper is used when the programmed configuration of the Control chip is
lost. In some cases, losing the configuration can cause the card to use addresses
that are already in use by other hardware. Changing the position of the Control
chip reset jumper from the two left pins to the two right pins disables certain
functions of the Gold card that could cause hardware conflicts. Once the jumper
is plugged on the two right pins, reconfigure the Gold card to the factory preset
values by issuing the following command:

setup /R

Once the Gold card is reconfigured, replug the Control chip reset jumper onto the
two left pins.

Note: You do not need to change the position of the Control
chip reset jumper to issue the command setup /R. Change
the position of the jumper only when the hardware conflicts
prevent the computer from rebooting. (See "Appendix A:
Troubleshooting", Problem No. 12, for details.)