Only 512mb ram? For shame... (sarcasm)

Id run NT if it was me, or 2K if I was lazy. I wasn't expecting XP. I guess it is a requirement if the video card drivers?

luckybob wrote on 2026-09-03, 13:11:

Only 512mb ram? For shame... (sarcasm)

I wanted 1GB damn it! 😀

luckybob wrote on 2026-09-03, 13:11:

Id run NT if it was me, or 2K if I was lazy. I wasn't expecting XP. I guess it is a requirement if the video card drivers?

XP is for the graphics card drivers, and I also prefer it to its predecessors.

If you run XP with the classic look, with no frills, it actually runs pretty well on systems such as this one.

I’ve already run XP on it, that’s how I got the Blender performance result for the machine.

I've purchased this NEC MultiSpin 6Xe CDR-602 External SCSI 6x CD-ROM Drive Caddy S/PDIF, which is tested and working as well as this Pioneer DVR-S201 External SCSI DVD-R Authoring Drive DVRS201 for this system: (So I'm getting there slowly)

Both of these drives will be removed from their external enclosures and slotted into drive bays in the system.

Here's some info on these drives:

The Pioneer DVR-S201 external SCSI DVD-R authoring drive was officially announced and released on June 22, 2000 and had the following specs:

The NEC MultiSpin 6Xe CDR-602 External SCSI 6x CD-ROM Drive Caddy S/PDIF has the following specs:

I've just ordered this 68-pin SCSI Removable Drive Rack to house the System Disk:

Here are its specs:

And I've also purchased this USB header, which will fit into one of the open parallel port openings on the rear of the case:

The NEC Multispin 6X is installed and working!

I've just purchased this ICP Vortex GDT8563RZ Ultra160 SCSI RAID Controller for the system, this will replace the ICP GDT6538RD 6x68pin, 4x50pin SCSI-Adapter that I currently have fitted: (It's the ICP GDT6538RD's rarer big brother.)

I've removed one of the IDE Drive Racks from the system to make way for an Epson SD-800 3.5" & 5.25" Combo Floppy Drive, as I need a system with USB support and Floppy Drive support as I have an 8086 and a 286 which I'll need to get files onto. (I also didn't like having two of the same units in the front of the machine!)

I don't need two IDE Drives in the system, one will be enough to do some testing!

So this thing's going to have 5.25"Floppy, 3.5"Floppy, CD, DVD, USB, Parallel, Serial and Network as IO options! Together with removable System and Storage Disks!

What do you plan on using this computer for because I think that CDROM drive is to slow to be useful it will seek allot.

Intel486dx33 wrote on 2026-10-05, 11:12:

What do you plan on using this computer for because I think that CDROM drive is to slow to be useful it will seek allot.

I'm not planning on using it for anything, once it's benchmarked it will sit on a shelf like all of my other systems. (My aim here is to combine as much interesting hardware as possible into this single build.)

This CD-ROM as well as this DVD-Writer are slow, but they're rare and super cool. (The DVD-Writer being the first commercial DVD-Writer, with an extremely cool look, and initial selling price of more than $4000, and the CD-ROM being a super cool implementation of the Caddy System.)

Here's the current list of parts, where I purchased them and for how much: (The prices include shipping)

The aim is to use as many collectable parts as possible. (Sometimes quirkiness outdoes raw performance!)

While taking a closer look at the picture of the ICP Vortex GDT8563RZ Ultra160 SCSI RAID Controller which I've just purchased, I noticed a critical missing component, without which I wont be able to boot off the card! (The optional BIOS ROM)

So I did some digging and found the BIOS update utility for this card and found that the BIOS is compatible with another ICP card in my possession: (The GDT6113RS!)

So it looks like I'll be able to take the optional BIOS ROM off this card and slot it into the GDT8563RZ, allowing me to boot off the new card!

As a final piece to the puzzle, I've purchased this NOS TEAC FD-505 Dual 3.5 & 5.25 Floppy Drive for the machine: (It's a much more solid option than the Epson)

The GDT8563RZ has arrived and it looks band new!

For now I’ve just tested it in my Pentium II Overdrive, and it’s working with the BIOS ROM that I took from the GDT6113RS!

This is what it shows as the card boots:

I finally have the answer as to what's been causing the instability in the Dual Pentium Pro System: (Bad caps on the CPU card!!!)

This morning one of the electrolytic caps in this region of the board gave up the magic smoke, with the associated fishy smell. I saw where the smoke came from and could smell the fishy smell in this area after I had pulled the card. (Not much smoke was given off, there is no leaked electrolyte, and none of the caps are bulging.)

So this thing's going to need a re-cap before I go any further! 🙁

There are two types of electrolytic caps on this card, but both types have the same value: 1000MFD 6.3V 105 Degrees Centigrade:

Can someone please explain what the difference between these two types of cap is? (As the one type is considerably taller than the other.)

I've ordered 40 Panasonic EEUFR0J102's to replace all of the electrolytic's in this system.

The base board also has five of them which will also need to be replaced:

The big green Sanyos on the CPU card are ultra low-ESR VRM caps

The panasonic FR series will probably be adequate substitutes

maxtherabbit wrote on 2026-10-17, 13:08:

The big green Sanyos on the CPU card are ultra low-ESR VRM caps

The panasonic FR series will probably be adequate substitutes

Thanks for the info, I really appreciate it! 😀

I just checked and got the following information: (I liked it up using 5V as a reference for a worst case scenario.)

Here's the voltage range that they should be dealing with:

The VRM in my system is currently set to 3.3 Volts.

I just ordered 10 of these EEUFR1C102LB's to replace the VRM caps, they're more substantial than the EEUFR0J102's (I'm doing this just for good measure)

Boy do I have a story for you today!

I started taking the pentium Pro apart for the recap and when removing the CPU's from their sockets I noticed that each of them was missing a grounding pin in the same place:

I was able to recover both pins and will ask the avionics at work next week if they can maybe solder them back. (I have another Pentium Pro which is in perfect condition that I'll use to test the machine in the mean time of necessary.)

But it still got worse than this!

When I was removing the base board from the machine, I noticed that I couldn't remove the ATX power coupling. On closer inspection I noticed that they had partially melted! (The only way in which I could get the board out of the machine was to cut the power couplings)

The underside of the base board doesn't show any scorch marks, which is a good thing!

I measured the 5V and 12v rails on the power supply under load, and they're fine:

The bad connection on the power coupling's more than likely the problem after all. I wasn't smelling exploding caps, but I was smelling the melting power coupling.

So I went off to the electronics store and purchased a new MOLEX KK 396 12-pin connector to replace the damaged power coupling:

I had to purchase a crimping tool too to fit the female connector to the power supply:

The power supply's rewired:

I'll have the male connector fitted when the boards are re capped.

Here are the specs on the MOLEX KK 396 12-pin connector:

So it can handle 35 watts per pin @ 5V, and there are four 5V pins, so that's 140W, which should be enough!

Here's a pic of someone else's board and the same pins are burned! (These things should have been fitted with 24-pin power couplings!!!)

@legodude on SGUG Discord suggested that I add some extra MOLEX power to the underside of the base board:

This is probably a good idea!

The thing about the obviously overheated connector is that the 35W of specified load is when new and spring force is still adequate. 30 year old stuff tends to lose spring force, especially if it's been disassembled and assembled several times. The resistance in the connection doesn't need to go up much before heat will be made - and it's a positive feedback loop just running away.
I definitely agree with you that this is the culprit for many instabilities. Congrats on sorting that out!