Hi vrob98,

Today I just managed to rescue one of those computers from the recycler. Researching my machine I bumped in your post. I don't think that the 27c is the bios. If you check your photo, you will notice that near the 27c there is a plcc32 ic (IC47) half covered by a sticker, with the typical Intel Flash logo. So I wage that under the sticker you will find a 28F010 or 28F001.
My machine appears fully functional, I can dump my flash as soon as this weekend.

A 1024K Flash chip for BIOS in a 386SX machine? I don´t think so.

TO also noted: "the board will succesfully POST and enter BIOS if I replace the original BIOS with a "MrBIOS" universal image for the same Tact83000 chipset"

Hi all,

There is nothing like checking, and lo and behold, a 28F010.

That may explain why there was a sudden malfunction. I have in my collection a non trivial amount of motherboards with failed 28Fxxx. Probably bit rotten, because after re-flashing they do work.
Now I'm somewhat concern because there is no available dump, and the size may point to some boot-loader code that is common on intel MB (or other proprietary shenanigans). Additionally I think the IC53 (pal or gal) probably has something to do with selecting portions of flash to something else than BIOS.

Any way I now feel some urgency to dump whatever is in mine specimen.

8KB contents on the dip eeprom sounds suspiciously like VGA rom, but quick glance and sure enough it starts like normal PC bios, gonna add it to my bios dissasembly zoo

Just dumped the 27C256B, its a match.

Made some dumps with ATHROMID. Its even stranger.....

Manage to remove the flash and made a backup dump with my T48:

Hope that it may be useful.

I think DIP eeprom is just a bootblock

1seg000:F020                 db  42h ; B
2seg000:F021                 db  6Fh ; o
3seg000:F022                 db  6Fh ; o
4seg000:F023                 db  74h ; t
5seg000:F024                 db  42h ; B
6seg000:F025                 db  6Ch ; l
7seg000:F026                 db  6Bh ; k

but why such elaborate setup 😮 ? neato

Just added a socket In the plcc footprint for future proofing (at least the 28F010). Put it all together and it works as expected. Also managed to archive the HDD, a Seagate ST3144A. Surprisingly ddrescue reported no faulty sectors.
As I suspected the small plcc is a PAL16L8 from Ti, by the traces there is a good probability that is related to this odd configuration.

I can now rest easy with this machine. Also the version of the bios is:

I went back to OPs

vrob98 wrote on 2026-03-27, 17:14:

ISA Analyzer shows 00/--, LPT Analyzer shows 77/6F

and sure enough Xicor you are right on the money!

1seg000:C560                 mov     dx, 378h
2seg000:C563                 mov     al, 77h
3seg000:C565                 out     dx, al
4seg000:C566                 sti
5seg000:C567                 call    RTC_RegC_IO_120_121
6seg000:C56A                 call    PIC1_enable_FDD_irq
7seg000:C56D                 call    FDD_Load_Drive_Parameters
8seg000:C570                 mov     dx, 378h
9seg000:C573                 mov     al, 6Fh
10seg000:C575                 out     dx, al
11seg000:C576                 call    FDD_read_firmware

vrob98s board has most likely (havent disassembled whole thing yet) detected corrupted flash and is trying to bootstrap itself by loading firmware for most likely (havent disassembled past that point ether hehe) rescue flashing. Whats more there is tons of routines beeping different error codes meaning vrob98 either has broken or unconnected PCspeaker 😀 because its beeping like crazy during this routine
Edit1: missed this line
>endless loop of 4 short beeps and 1 long beep

so it is beeping like crazy 😀

Edit2: and it is indeed beeping because its failing to find a floppy with data for recovery.
FDD_read_firmware starts by calling FDD_read_track0_sector1 (all names invented by me)
FDD_read_track0_sector1 as the name implies tries to read track 0 sector 1 twice, if both tries fail it does this

1seg000:D446                 mov     ax, 1
2seg000:D449                 call    BEEP_error_code
3seg000:D44C                 mov     ax, 500
4seg000:D44F                 call    Delay_AX_milliseconds
5seg000:D452                 jmp     short read_sector1_loop

and the Beeper function:

1seg000:CB85 BEEP_error_code proc near
2seg000:CB85
3seg000:CB85 error_bitmask   = word ptr -2
4seg000:CB85
5seg000:CB85                 push    bx
6seg000:CB86                 push    cx
7seg000:CB87                 push    dx
8seg000:CB88                 push    bp
9seg000:CB89                 mov     bp, sp
10seg000:CB8B                 push    ax
11seg000:CB8C                 mov     bx, 4
12seg000:CB8F
13seg000:CB8F BEEP_loop:
14seg000:CB8F                 mov     cl, bl
15seg000:CB91                 mov     ax, 1
16seg000:CB94                 shl     ax, cl
17seg000:CB96                 test    [bp+error_bitmask], ax
18seg000:CB99                 jz      short BEEP_1355Hz_120ms
19seg000:CB9B                 mov     ax, 400
20seg000:CB9E                 mov     dx, 784         ; 1520Hz
21seg000:CBA1                 jmp     short BEEP_it
22seg000:CBA3 ; ---------------------------------------------------------------------------
23seg000:CBA3
24seg000:CBA3 BEEP_1355Hz_120ms:
25seg000:CBA3                 mov     ax, 120
26seg000:CBA6                 mov     dx, 880
27seg000:CBA9
28seg000:CBA9 BEEP_it:
29seg000:CBA9                 call    BEEP_AX_milliseconds_DX_frequency
30seg000:CBAC                 mov     ax, 120
31seg000:CBAF                 call    Delay_AX_milliseconds
32seg000:CBB2                 mov     ax, bx
33seg000:CBB4                 dec     bx
34seg000:CBB5                 test    ax, ax
35seg000:CBB7                 jnz     short BEEP_loop
36seg000:CBB9
37seg000:CBB9 Halt_here_loop:
38seg000:CBB9                 test    byte ptr [bp+error_bitmask], 10000000b
39seg000:CBBD                 jnz     short Halt_here_loop
40seg000:CBBF                 leave
41seg000:CBC0
42seg000:CBC0 end:
43seg000:CBC0                 pop     dx
44seg000:CBC1                 pop     cx
45seg000:CBC2                 pop     bx
46seg000:CBC3                 retn
47seg000:CBC3 BEEP_error_code endp

ax = 1 results in four short beeps (1350 Hz, 120 ms) followed by one long (1520Hz, 400ms). Then 500ms delay and it tries reading sector 1 again. If vrob98 had Floppy 0 connected he could also notice motor turning on, Floppy wouldnt be making any head movement/noises as we are repeatedly trying same track.

if im mega super bored I might try reverse engineering what format of data its expecting from the floppy

vrob98 if you ever come back or someone else happens to have deteriorated Intel flash chip on this board:
1 format standard 720KB or 1.4MB floppy. Only 9 and 18 sectors per track supported.
2 download download/file.php?id=249990 graciously uploaded by Xicor
3 extract VGA_BIOS.BIN and rename to VGA74_vv.LDB or VGA74_.LDB
4 extract BIOS_ROM.BIN and rename to B22_vvvv.LDB or B22_.LDB
sidenote: Anything in filenames between '_' and '.' is ignored, guessign its their magic wildcard so can still put version numbers in filenames like VGA74_01.LDB VGA74_03.LDB etc.
5 Put those two files in the root directory of a floppy
6 boot with floppy in drive plugged/connected as A/first.

There will be some short beeps followed by flashing firmware to Intel chip.

Outstanding job rasz_pl !!
I will try to find in my stash a 28F010 and test. It boggles the mind why ICL would feel the need to have a somewhat advanced feature in a very low entry level computer, worst yet my machine is from 1993-02-02, extraordinarily late for the 386 family.