SouthMissDJ wrote:

I recently found this motherboard in a system being thrown away. I took this photo to document all of the jumpers before giving the board a bath. The settings pictured are for the Intel 486DX2-66. I have since upgraded CPUs to the AMD 5X86-133, an undocumented processor for this board. The closest settings get it to either 120 or 150 mhz. Good luck with yours!

what jumper setting did you use to set it at 120mhz or 150mhz
(im talking about an AMDX5 133mhz cpu)

because that is not tocumented anywhere.

Hi
What should go there in RN13, RN14, RN16?
Is it some resistors leader?

RN=resistor network

PcBytes wrote on 2022-08-09, 21:51:

RN=resistor network

Ok, but could You measure values?

I don't have the board but figured it would be one of those SIPP styled resistor networks.

Although, most of the photos I see over Google show them as being left empty, except this one which appears to have what I think is a dummy resistor network (000J means 0 ohm, I guess?) so probably those sockets are unused altogether.

Yes FIC used 0 ohm arrays instead of jumpers on many 486 boards, somewhere here is a post about them showing 0 ohms and their layout.
Basically there are 5 loops of wire inside pins 1-2, 3-4, 5-6, 7-8,9-10.....
like in this topic: Thoughts on how to go about fixing this?

Horun wrote on 2022-08-09, 22:45:

Yes FIC used 0 ohm arrays instead of jumpers on many 486 boards, somewhere here is a post about them showing 0 ohms and their layout.
Basically there are 5 loops of wire inside pins 1-2, 3-4, 5-6, 7-8,9-10.....
like in this topic: Thoughts on how to go about fixing this?

Thanks, That would be easy to make.

DeafPK wrote on 2018-02-09, 13:03:

I am currently working on one of these mobos - it goes through its POST, it can auto detect a 500 meg HDD over a Goldstar ISA I/ […]
Show full quote

I am currently working on one of these mobos - it goes through its POST, it can auto detect a 500 meg HDD over a Goldstar ISA I/O controller, but it won't boot. Not from anything. The floppy does not search for disks after POST but it reacts during the POST. CMOS battery has been replaced with CR2032 mod and it works.

What happens is POST does its things, seemingly successful, then the screen clears and there is just a single flashing underscore in the top of the screen, like _
I gave it 15 minutes once, no gain. The stuff connected to the system are parts known to be OK. I have to say that the mobo was pulled from a dumpster and I've cleaned it with isopropanol alcohol and had it drying for months. Who knows really what is has been through.. I just don't feel like giving it up when I'm getting so close.

BTW if anyone still wonders, J7 gives a message before POST that BIOS passwords are disabled.

I know it's been a while but just to add some clues for future readers: I had the *exact* same setup and symptoms, and eventually found the ISA FDD/HDD controller to be the culprit. Exchanging it for a Winbond ISA controller card solved the issue on the spot.

Finally the full user manual for 486-GVT-2 is available on TRW site. Unfortunately it contains no descriptions of the jumper settings for any of the AMD 486 CPUs.

I'm also looking how to setup jumpers properly for AMD - the Am5x86. I have one at interposer with onboard VRM so voltage is not problem. With latest BIOS it's even detected correctly but for some reason it hangs at POST code 61h when L2 cache is enabled in BIOS. With L2 disabled I can boot to DOS but performance seems to be poor compared to others MBs...
BTW I had RN16 in position and I can confirm it contains just 5 jumper wires inside like this:
*-* *-* *-* *-* *-*

Aha, I found the problem with L2 cache is only at FSB 40MHz and higher, for 33MHz and less it works fine. I populated all 9 chips 32kB with 15ns access time, maybe do I need 12ns tag RAM to work at 40MHz? But I wonder it doesn't react on cache speed setting in SETUP (normal/fast/turbo), at 33MHz it works in turbo mode and at 40MHz it hangs at any mode. This options are a bit cryptic instead usual 0WS/1WS etc...
Also I found that option for Alter and Tag bit combine speed it a bit up. Interesting that when I switched L2 from WB to WT it hangs even at 33MHz (still POST code 61h).
I tried also 50MHz FSB with L2 disabled and it boots but I got VGA errors (messed palette in Doom) on Trio64V+ VLB, regardless those 2 VLB jumpers - seems they have no any performance effect. So 50MHz is no-go. Problem is that 5x86 runs with 3x multiplier only. I tried to flip every jumper on the MB and was unable to switch it to 4x multiplier. Maybe the CLKMUL pin (R17) on the CPU socket is not wired anywhere (on 486DX2 it's marked as not connected, introduced on DX4) so I should tie it to GND manually...

I traced the CLKMUL (R17) CPU ping goes to pin 4 of resistor network RN16 which connects it to pin 191 of VT82C486A chipset (no datasheet). It seems it was intended for Cyrix/IBM CPU that has signal HITM# at pin R17 instead.
So I used a piece of wire to connect RN16 pin 4 to GND and enabled the 4x multiplier.

Unfortunatelly still no luck with L2 cache at 40MHz FSB - hangs with any BIOS settings (POST code 61h).
I tried to replace TAG chip by 12ns but no change. Without L2 the Doom score is only by 0,5FPS better than at 4*33MHz with enabled L2 WB

Also wondering than manual shows insane FSB speeds 50-100MHz, of course it doesn't work...

UPDATE: Accodring to SpeedSys and benchmark results, the L1 cache of Am5x86 always run in WT mode regardless BIOS and jumper settings. I tried toggling all related jumpers but unable to set WB mode. So this MB is definitely not the option for high-end 486 CPUs...

UPDATE2: I traced the WB/WT# (B13) CPU ping goes to pin 6 of resistor network RN16 which connects it to output pin 12 of U19 PALCE16V8 prog. chip. It seems it was intended for Cyrix/IBM CPU that has signal WM_RST at pin B13 instead.
So I used a 1k resistor to connect RN16 pin 6 to +5V (from near CB42 pin) and enabled L1 WB mode. It resulted in significant speed up but it's still behind my top VLB Octek Hippo 10 MB...