VOGONS


First post, by cluster.fsck

User metadata
Rank Newbie
Rank
Newbie

I inherited a 486DX-33 system at work. UMC 481/482 chipset, 256KB L2 cache, AMI HiFlex BIOS 07/07/91 core. BIOS ID string is 40-0100-ZZ1211-00101111-070791-U480A/2-F. Looks to be this board https://stason.org/TULARC/pc/motherboards/D/D … 6DX2-50UCE.html. I'd like to max out the RAM, but am I correct in assuming the 32-bit memory expansion slot is proprietary to this motherboard and chances of locating an expansion card near-zero?

I also see that this chipset supports up to 1024KB L2 cache, but the board is only wired for 256KB. Would it be possible to mod this board to accept 1024KB L2 cache? I guess a better question would be is it worth it? My gut's telling me the performance impact would be negligible, but the "wasted potential" still irks me. I successfully performed a PS/2 mouse mod inspired by the thread on this board, so maybe I should quit while I'm ahead. I'll post details and a BIOS image in a few days.

Thanks.

Attachments

Reply 1 of 2, by root42

User metadata
Rank l33t
Rank
l33t

Don't fix a running system...

That said I have a mysterious 386 board with a similar expansion slot. Never figured out what it's for. RAM or Cache... I never owned a 486 with more than 256 KiB cache. Would be intriguing, but I guess the limiting factor on this machine will be the ISA bus.

Edit: this is my board

8F7BEE58-B2FE-4511-9950-9949C67DCAF5.jpeg
Filename
8F7BEE58-B2FE-4511-9950-9949C67DCAF5.jpeg
File size
270.09 KiB
Views
355 views
File license
Fair use/fair dealing exception

So Opti vs UMC and you have two long edge connectors for the expansion, while mine has something looking like a 16bit ISA.

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 2 of 2, by cluster.fsck

User metadata
Rank Newbie
Rank
Newbie
root42 wrote:

Don't fix a running system...

Yeah, but 1024KB L2 cache would be sweeeeeetttt.....

In regards to the expansion slot, I killed a PC CHIPS 386 board by tracing through it's proprietary slot with a multimeter. Now it beeps saying memory error in lower 64KB. Not gonna repeat that mistake with this board. Guess those old CMOS ASICs are as fragile as they say 😢

I've attached the BIOS image I dumped with a device programmer. I dropped the image into IDA Freeware and found the checksum routine- the 16-bit sum of all 16-bit little-endian WORDs must be 0000h. To verify, I swapped characters of WAIT to ITWA and the board still boots. I'll do a little writeup on some other things I found when I get a chance.

Attachments