Whats written on the pcb near the VRM? LT1083? or is it LT1085 or 1086?

I have that board and I modded it to support 3.3V with parts I had at hand but the board is still not all that great and doesn't seem to work stably at 40MHz or higher bus speeds.

Regulator I used is NCP1086-33 and the mods are primarly to allow more current to reach the CPU since everything seems to go through single jumper.
JP12 is 3.3V / 5V jumper and I had to remove JP9 which purpose I no longer remember.

I have placed EZ1084CM, adjust.
My run with DX-50 stable at 50Mhz FSB. but not with DX4, only 2x Multi.
JP9 is bypassing to 5V or for 3.3V over resistor at 5V. I have removed solder jumper.

I think: place 3,3V CPU Support at this board is money & time-waste. 😒

Can you guys make a nice photo for here and dump the ami bios on your boards ?
http://www.win3x.org/uh19/motherboard/show/73

Thank you 😁

My board has an Award BIOS, which I attached. Board itself has been slightly cannibalized and a photo of it probably isn't the best.

Mine has Ami-WINBios.

I haven't good camera (chinesium-smartphone)

My Photo, no high quality.

You're allowed to take of this pictures to this side.

I give up, i'll selling this board for price (26.90€) , what i have paid. I don't want keeping this board.

Problem solved ! My multi IO VLB and CF-IDE Adapter is reason for instable DX4-run. Reason: I am lazy and i let run CF-IDE adapter without external 5V supply. (CF IDE adapter run on 5v from IDE Port) Abit board and one vlb ide card doesn't like this way. I plug of 5V supply on CF IDE adapter and DX4 run fine.

My Board has rev with four jumper for voltage selection.
Modification-List is only this Rev: 2:1.1

Q4 = LT1085-ADJ or pincompatible type with LowDropout(LD1085V from ST, EZ1084, ...)
TC15 = 10µF capacitor (eletorlytics or Tantal,whatever!)
TC12 = 150µF eletrolytics (recommend by LT1085 datasheet), i use 470µF
JP52, bridged with wire by soldering.
R77 = 270 ohm , 4V CPU voltage , Jumper JP53
R78 = 240 Ohm , 3,6V CPU voltage, Jumper JP54
R79 = 220 Ohm, 3,45V CPU voltage, Jumper JP55
R80 = 200 Ohm, für 3,3V CPU voltage, Jumper JP56
R81 = 127 ohm
R41 and R42 is placed with 10 Ohm (my estimated value ) Location: between VLB slot and CPU socket
All eight capacitor in CPU socket is 100-470nf capacitor.

And the END: JP9 is recommend to setting,
Intel DX4 need this jumper JP9, 1-2 or 2-3, it is not important (recommend (1-2 )). It is possible, it run without jumper stable. I
But AMD DX4 doesn't need him, because this PIN is N.C. It is for 5V tolerance I /O buffer of Intel DX4 and 5v Supply of 486 with 5V

My BIOS doesn't support AMD DX4 and 5x86. it is too old. 😒

the AH4 seems to share the same bios with the AN4R2. I have version 4.02 on file

108/30/95-SIS-85C471E/G-2C4I9A12-02
2GREEN 486 VESA SYSTEM BIOS

If you have the possibility to flash back the old bios externally if this one does not work properly, you can try it.

Release notes:

1Date: 8/31/1995
2ID: 02
3Enhanced to support Intel P24D, Cyric M1sc, DX4, and 120MHz CPUs.  

You can probably download the latest Award BIOS from august 1995 from the UH19 page. I'm sure it'll support AMD 5x86

Edit : oh well evasive got quicker x)

Yes, this BIOS image is GOOD 😀

I have pile Intel 27512. I pick one and burn image.
Yeah, AMD 5x86 run, (detected as (enhanced )AmDX4-S or similar)

I can tell: Overclocked Intel DX4 (100@120 ) is a true beast, eaten AMD 5x86-160 in few bench .

Thanks
Matt

Next success step.

ABIT AN4R2 and AN4TR2 doesn't support writeback, becaucse: PCB has a design fail.
WB/WT-Pin of 486 is connect to empty solder-joint for jumper. Fail: Jumper short WB/WT Pin to ground. It should connect to VCC, if WB Mode is needed.

And it run faster 😀

Modification: JP 59, 37, 63 is empty solder joint, added with pin header

Jumper setting for WB Mode
JP59 (2-3 )
JP 37 4-5 (Intel DX4 works in all combination, whatever 1-2, 2-3, 3-4 or 4-5 , wreid , but AMD 5x86 doesn't like 2-3)
JP35 1-2
JP63 <- This is reason, why ABIT AN4R2 doesn't support WB, design fail. (JP63-Pin 2 should connect to 5V, here : Ground )
JP63 Pin 1 connect to 5V, if WB Mode is needed. (Recommend over 1k Resistor )

EDIT (15.12.2020) JP37 is not neccesary, open jumper -> no influence about WB-Mode, but other two jumper has influence.

Result :

(512kB L2 Cache is a chinese counterfeit of ISSI 1MBit SRAM, this is why, i am going back with L2 Cache timing)