Alright, I guess I won't pack the system away just yet.

1label:
2mov  CX, 1000
3myloop:
4loop myloop
5
6mov	DX, 0x80
7in      al, 0x64
8in      al, 0x60
9
10cmp al, 0x29
11jne dalej
12mov al, 1
13dalej:
14cmp al, 0x10
15jnb label
16
17dec al
18out	DX, AL
19
20mov  cx, 1000
21myloop2:
22loop myloop2
23
24mov     bx, 0x1000
25mov     es, bx
26xor     si, si
27and     ax, 0xf
28add     si, ax
29mov     ax, es:[si]
30out	DX, AL
31jmp label

pressing "~1234567890-=<backspace>Tab" keys will output one byte + index. Every key is one value from 0 to 14 (oops missed one)
Should be AA 55 5A A5 CC 33 3C 3C 55 AA A5 5A 33 CC C3

I turned on the motherboard with your latest BIOS mod installed. Screen stays blank. I installed a POST card in the ISA slot, but the POST card flashes garbage. I think it is trying to write ErEr. Typing in series ~1234567890-=<backspace><tab> didn't change the value shown on the POST card, nor did anything show up on the monitor.

EDIT:
Do you think I should bother tracing out all the PBSRAM data pins to ensure they go to the 82438FX TDP chip?

I screwed something up, It worked fine, and now doesnt, I think I might have packaged and send wrong file while deleting working one, rechecking 🙁

Edit: nah, it works, just weirdly. I think there is something wrong with my code and im slamming post card too fast 😐 try this slowed down version, if not Ill rewrite it with more proper slowdown routine
edit2: added moar delay loops, not its ~400K cycles per loop so ~500Hz on Pentium 200MHz? might still be too fast
edit3: ok, now about 2-3 updates per second
edit4: you know its good when its the 4th version of the second version! now ~40 keys work, Q-P []\ A-L ;' Enter etcetera. I thought it would be prudent to look at more than 16 bytes just to be sure.

feipoa wrote on 2022-11-17, 12:47:

Full Yes 82430I PCI Bus Main Board: https://theretroweb.com/motherboards/s/full-y … 430fx#downloads

There's a guy on youtube with this board, https://www.youtube.com/watch?v=CEiSQsDI8dQ , but he is only installing the 512K asynchronous cache. He does point out the jumper to set to PB SRAM though.Anyone know if he is a user here? I don't see the nick vswitchzero on Vogons. I'd like to see if he can trace where JP35 goes on his motherboard so that I can replicate it.

Hi feipoa,

My handle here is lemonlime. Would be nice to change it to vswitchzero but don't see that option. If anyone knows how I can request it to be changed, please let me know 😀

I just took a close look at the Full Yes 82430I under good lighting. I had it in my head that it was the two 3-position jumpers (JP1 and JP2) hardwired near the PB positions that were used to switch the cache type, but I was wrong about that. Those two are for setting 5V/3.3V for the cache according to the manual. I stared at it for ages, but I can't find anything labeled JP35 anywhere on the board. All of the jumper banks I can see are accounted for in the manual, including a few hardwired ones. The highest numbered bank I could find was JP33/34 near the I/O ports. The scan quality of the manual on the retro web is really poor, so I can't see the documented positions, but it seems to indicate that there are supposed to be three 3-position jumpers in the spot where JP1/JP2 are. I only see two there. But it seems strange to me that they would have a PCB revision without the jumper, but still bothered to include the PB locations and traces. Very strange.

I'm very curious to see what you are able to find, because I was hoping to do a similar mod on this board as well. Looks like it won't be as simple as I had hoped.

Here are some higher-res photos of the board. I'm happy to check or trace anything that may be helpful to you. Just let me know.

rasz_pl, I'll try the new BIOS mod soon. To confirm, I'm to read numbers off the POST card and the monitor will be blank?

lemonlime, I don't think there's a way to change your username, otherthan creating a new username.

Thanks for checking your board. It is unfortunate you didn't have the version of the Full Yes board with JP35. I didn't finding a listing in the manual for what JP33 and JP34 are for. Maybe on your revision on the board, these jumpers are for ASYNC/PBSRAM?

It is interesting to see how your board uses 4 ceramic caps + 1 tantalum per PBSRAM pad.

If JP35 does exist on your board, it may very well either be a soldered wire between vias, or even a trace connecting two vias. I spent 10 minutes looking at the photos, but didn't find JP35. Under the ASYNC SRAM sockets maybe? lol

rasz_pl, I tried the new BIOS.

I see on the POST card, without pressing any keys:

It stays like that for a long time, but after about 5 minutes, I see:

Pressing keys doesn't do anything.

feipoa wrote on 2022-11-25, 19:54:

lemonlime, I don't think there's a way to change your username, otherthan creating a new username.

An admin should be able to change this via SQL or in the backend maybe. Maybe write them directly via "contact us"

feipoa wrote on 2022-11-25, 20:25:

rasz_pl, I tried the new BIOS.

I see on the POST card, without pressing any keys:
POST_card_immediate.JPG

all good

feipoa wrote on 2022-11-25, 20:25:

It stays like that for a long time, but after about 5 minutes, I see:
POST_card_after_5min.JPG
Pressing keys doesn't do anything.

this is extremely weird :0 must be POST card internal logic, maybe times out

I was too smart ass, emulator is not real hardware after all. I tried reading keyboard before keyboard controller was initialized and enabled (post code 05). One more try, Ill try calling 05 routine first.

I thought Asus firmware is same for this board for either pipelined cache or async cache unless this custom board was created solely for one maker?

Had you have installed '245 buffers yet? Does the silkscreened on the board itself have jumpers that shows pipelined cache and async cache configurations, then follow this jumpers set up for pipelined cache.
Also, Mix is for 5V to use async cache only. When using pipelined cache, must be set to 3.3V only.

Cheers,

change of tactics, initializing keyboard controller was too much work 😀

1mov  bx, 0x1000
2mov  es, bx
3xor  si, si
4label:
5mov  ax, 0x100
6xor  cx, cx
7myloop:
8loop myloop
9sub  ax,1
10jnb  myloop
11
12mov  ax, si
13mov	 DX, 0x80
14out	 DX, AL
15
16mov  cx, 0xffff
17myloop20:
18loop myloop20
19
20mov  ax, es:[si]
21out	 DX, AL
22inc  si
23cmp  si, 0x10
24jb  label
25xor  si, si
26jmp  label

this will just loop over first 16 bytes, its almost too fast with emulated 200MHz cpu, its ~1 update per second with 75MHz one. If its too fast just record video of post card with your phone and watch it back in slomo.
Correct output is "AA 55 5A A5 CC 33 3C 3C 55 AA A5 5A 33 CC C3"

The POST card continues to show C1|00. Pressing in series~1234567890-=<backspace><tab> does nothing. Keep in mind that on on western keyboards, producing ~ requires a <shift>`

Nonetheless, I tried both ~1234567890-=<backspace><tab> and `1234567890-=<backspace><tab>

The last one didnt use keyboard at all, no need to press anything. It should sequentially scroll bytes ~1 per second on the POST card.
Are you sure you flashed the last one? 't15i0302 lame sequential mod.7z' that would meant your board hangs after line 15 Re: Anyone own an Asus PCI/I-P54TP4 with pipeline burst SRAM? Planning to mod direct-mapped SRAM to pipelineburst but there is nothing there to hang at.

I can rewrite it using different registers, but I would think emulating CPU is one thing x86 emulator got right. There is nothing else here. Just cpu and POST card outputs, with generous delays in between.

1mov  dx, 0x80
2mov  al, 0xff
3out  dx, al

output ff as a marker so you can be sure right file was flashed

1mov  bx, 0x1000
2mov  ds, bx

segment register to where cache_test stored magic sequence

1start:
2xor  bx, bx
3mainloop:
4mov  ax, 0x100
5xor  cx, cx
6delay:
7loop delay
8sub  ax, 1
9jae  delay

big delay loop, 256x65536 cycles

1mov  al, bl
2out  dx, al

index in bl, output index

1xor  cx, cx
2delay2:
3loop delay2

small delay

1mov  al, byte ptr [bx]

ds:[bx] -> 0x1000:0-15

1out  dx, al
2inc  bx
3cmp  bx, 0x10
4jb   mainloop
5jmp  start

It would be unlikely that I burned the incorrect ROM. I renamed your file so I know the correct ordering of events. See screenshot. But to be sure, I burned it again and booted.

The latest ROM shows FF|C1, however I can tell it is scrolling through some other codes, but WAY to fast for even 30 fps to capture. You might need to increase the delay 10-fold.

feipoa wrote on 2022-11-27, 01:44:

The latest ROM shows FF|C1, however I can tell it is scrolling through some other codes

if it stops at FF then its not showing other codes 🙁
nevetheless, 16x and 256x slower
Previous delay of ~16mil cycles would put it at ~10 updates per second IF you used 200MHz Pentium level cpu capable of running 1 instruction per second, and afaik those dont exist, unless you are testing on K6 480MHz (this boards maximum?) to troll me 😜

if this doesnt work Ill just make 8 bioses displaying 2 bytes at a time 😀

Will check later.

I'm using a Pentium 133 overdrive.

overdrive in socket 5?

lol. It is not a literal overdrive, but it looks like an overdrive, with its integrated fan/heatsink. Photo posted earlier, download/file.php?id=150307&mode=view

vogons attachment server died/was turned off

Tried both images in t15i0302 oso looong11.7z

Both still show FF.