VOGONS


First post, by Eep386

User metadata
Rank Member
Rank
Member

Got this well beaten-up board on eBay. Typical problems: considerable VARTA battery damage and what looks like rodent "effects" (yuck!). Cleaned the board and used some magnet wire to jump some traces. However a 74F00 was in pretty bad shape (corroded inside and out!) and one of the pads lifted as I attempted to desolder it. So I had to resort to a really ugly flying wire to get it working again after I replaced the 74F00. I'll try to make it look a little nicer with Kapton tape once I get some. I also replaced a couple of mouse urine-ruined ISA slots, but I still need to replace the 8-bit ISA slot. (I was able to get the rest of the slots to a sufficiently-working level with a brass brush, but I wasn't able to save all of them...)

I identified the chipset: The Morse 91A320/321 are really rebranded OPTi 82C391/392 chips. Testing with an MR BIOS for the OPTi 82C39x confirms this, board works great with MR BIOS and even scores slightly faster with it too (all of 3-4ns faster off-cache, whoop de doo).

Has socket for dirty TAG SRAM, but it requires a funky 24-pin 16Kx4 SRAM chip. Am using a Logic L7C165PC-20, seems to work great with it. At the original 33MHz the board came with (66MHz osc), it scored 114-ish off-cache without it under CACHECHK; installing dirty TAG dropped the access time to the low 70s. (Final speed with a 40MHz bus is mid-50s off-cache.) Maxes out at 128K cache, which is fine as it's enough for 16MB in write-back mode, though as there's only one bank of cache I suspect the lack of interleaving hurts overall cache efficiency a bit.

Only real gripe I have with this board, is it doesn't directly support Cyrix 486DLC CPUs (a limitation of the old OPTi 82C391/2 chipset), so it's stuck with an Am386DX-40 for the rest of its life.

Morse M3.jpg
Filename
Morse M3.jpg
File size
1.64 MiB
Views
362 views
File license
Public domain

The underside (sorry about that nasty flying wire... and even more hideous black tape!!)...

Ugly repair.jpeg
Filename
Ugly repair.jpeg
File size
1.57 MiB
Views
362 views
File license
Public domain

And the BIOSes... first is the original AMI BIOS, the second is a compatible MR BIOS. Use a 27C512 compatible if you want to burn your own chips.

Filename
Morse M3 AMIBIOS 05-05-91.zip
File size
36.93 KiB
Downloads
24 downloads
File license
Fair use/fair dealing exception
Filename
V020B324.zip
File size
47.68 KiB
Downloads
20 downloads
File license
Fair use/fair dealing exception

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 1 of 5, by Eep386

User metadata
Rank Member
Rank
Member

Aaand it broke.

Looks like the 82C206 (Morse 91A206, aka. OPTi/Samsung KS83C206) gave out. I'm really starting to detest these '206 chips...

While I'm waiting for the replacements to arrive (pray that they're good!), I fixed the nasty flying wire in the meantime.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 2 of 5, by Eep386

User metadata
Rank Member
Rank
Member

OK, fitted a new KS83C206 chip. One of the pads lifted, but *luckily* it makes contact with the pin. I don't think I really want to mess with the chip any further at this point.

After powering it up...
'-- --'.

... after the dust settles, I take a good look at the chip. Found some microscopic bridges between pins... cleared em out with a hobby knife. Retested, '-- --' again, but this time there was some activity on the #IRDY indicator on my POST probe. Ok... that's a little different. Pressed the '206 chip and repowered, same thing. Pressed on the FPU, and... wait, what? It booted!!?

Huh. Flexed the board *gently* as a test, and we were back to '-- --'.

Pressed on the CPU and repowered... it booted!

OK... so my 386 isn't making good contact with the board. Reflowed one side of the PQFP pins (it's one of those PQFP-on-PGA-board things) and one corner of the socket just for good measure and...

...It boots.

Flexed the board, still boots.

I ain't gonna celebrate just yet though. I'll keep it under observation for a while just to make sure this isn't a fluke. It's nice to see this damn thing show some life though, after all I had to go through.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 3 of 5, by weedeewee

User metadata
Rank l33t
Rank
l33t

Congratulations ! 😀

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 4 of 5, by Eep386

User metadata
Rank Member
Rank
Member

Thanks! After replacing the rodent "effect" damaged 8-bit ISA slot, I've got it all put together and running. So far it seems to be running okay, but there's one thing I noticed about this board: the ISA bus divider seems fixed at CLK2IN/3, and there's no way to change it in either AMI or MR BIOS.

So at 40MHz (80MHz osc) the bus was running at 13.3MHz... probably why the original 82C206 gave out, as originally it was running at 33MHz / 66MHz osc for 11MHz. So in the interest of preservation I lowered the system osc to 66MHz for a 33MHz CPU clock and 11MHz ISA clock. Here's to hoping that it runs well from here on out!

Here's the board in its current state. (Taken before I switched the oscillator.)

I hope I'm done with this.jpg
Filename
I hope I'm done with this.jpg
File size
1.23 MiB
Views
267 views
File license
Public domain

And no more ugly wire on the back! Whoo!

Fixed Back.jpg
Filename
Fixed Back.jpg
File size
788.78 KiB
Views
267 views
File license
Public domain

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 5 of 5, by Eep386

User metadata
Rank Member
Rank
Member

Another minor update... seems the board didn't really take as well to the MR BIOS I picked as I thought, weird graphical glitches started creeping in here and there in Windows, and Windows'd occasionally stop reading the mouse.
So I put it back on the original AMI BIOS, and am re-testing. This one seems to set a higher divider, CLK2IN\4, so I now get a much slower but now seemingly stable system 🙁 At least Windows hasn't reacted viciously to it yet...

Further update: AMI BIOS seems stable, even with 40MHz CPU clock. I'll test this for a while, if it passes my battery of tests I'll call it good.

Final update (hopefully): Nope, 40MHz isn't completely stable. HIMEM reported a RAM error... back to 33MHz it goes. Bummer. It seems rock-solid at 33MHz at least.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁