Hey, that's a great hint, thx! I'll try that on my PC-Chips M571 mobo which came which a 333GP (75MHz Version but running @83MHz).

And don't worry, I'll check the TNT in a UMC board. Is the GA486AM/S ok? I take this for testing because the HOT433 always is a bit of.. umm.. unrealiable you know it sometimes produces strange results that may or may not be reproducable. So the HOT433 is not a good board to verify if card A works with chipset B.

//edit: Answer: Yes, the TNT works in a UMC board. Tested with 3dbench and PCPbench.

Great. Thank you for confirming this!

Any idea if the TNT Windows 98 drivers work on a 486?

Are you guys interested in the manual for the 4DPS? revision 2.1, I could scan it an email if anyone is interested.

Is it any different than this PDF manual I found online? If so, yes, I'll take your manual.

No worries, its exactly the same.

After a long time of messing around with older systems and other stuff I got back to this one. I swapped the LG 52X CD-ROM drive for a Creative Infra 1800 (with remote) that a friend gave me recently, I think it fits a lot better than the 'modern' LG drive. It just needed a lens cleaning and it's good to go. 😁

Next I decided to try the 4.00 BIOS. Burned it to a 27C1001 using a standalone programmer, swapped the chip and cleared the CMOS by setting J6 to 3-4 for ~30 seconds as FGB recommended. It came up fine but I didn't get a CMOS Checksum Error or similar - in fact the CMOS settings were intact. When it gets to the boot stage it stops for a few seconds, then displays a white rectangle on the center of the screen and locks up hard. It doesn't even try to boot from a floppy or hard disk.

Please excuse the terrible picture. I tried setting the clear CMOS jumper for 10~20 minutes, settings still intact. Then I removed the battery altogether and went out for dinner hoping that it would clear itself, but when I got back (1~2 hrs) the CMOS settings were still intact! The CMOS data remanence on this motherboard is amazing.

I tried restoring BIOS defaults and toggling some settings to no avail, I get the white rectangle of death every time. For the time being I put the 1.6 BIOS back but now the mouse cursor has become jumpy on Windows 95, like some sort of random stuttering but only the mouse cursor seems to be affected. I'm gonna leave the CMOS battery out overnight in the hopes of getting at least the mouse issue sorted out. I feel like I should've stuck with "if it ain't broke, don't fix it". 😢

I have uploaded the 1.6 BIOS image dumped from my motherboard as requested by feipoa. 😉

Thank you for uploading BIOS version 1.6. I've added it to my archive.

What happens if you only use 1 stick of RAM, tried seperately in both slots?

I didn't have the white box issue when when using BIOS 4.00, but found 1.72f to work the best. What happens if you flash to 1.72f using the DOS flash software? Refer to attachment for AWARD flash and BIOS 1.72f.

FGB wrote:

I remember that I had to higher the RAS to CAS Delay fron 2 to 3 when running the 133@160.

Do you recall what other BIOS settings were optimal for 160 MHz?

EDIT: I thought it was the BIOS settings at 160 MHz causing issues, however it seems the 3.35 V given to the X5 at 160 MHz just doesn't cut it. 5 V works, but I fear it will shorten the CPU's lifespan. Some boards have a 4 V setting, but setting the jumpers there doesn't yield 4 V. Looks like I will need to modify this motherboard such that the 3.35 V setting is 3.65 V.

Cool boots, man!

For those who want to modify this motherboard (Zida Tomato 4DPS), I have enclosed some photos. You need to remove the purple-circled resistor and replace it with a 1 K-ohm trimmer. This lets you adjust the CPU voltage from 3 - 5 V.

While the fastest cache settings of 2-1-2 worked with a 33 MHz FSB at the stock 3.35 V, no chipset/cache/memory settings would work at 40 MHz at 3.35 V with an X5 ADZ at 160 Mhz. Even at 3.7 V, I had to use 3-1-2 cache timings.

The interesting discovery here is that increasing the CPU voltage further allowed the cache to work at the fastest setting of 2-1-2. I thought the cache settings would have been independent of the CPU voltage. What's going on here?

At any rate, 3.95 V seems to be the minimum stable ADZ voltage if you want to use 2-1-2 at 40 MHz. I did just get a BSOD after playing an mp3 on repeat, so I'll try upping the voltage further.

I hope someone finds this information helpful.

What is the maximum voltage you can run an X5-133ADZ for continual operation without the risk of damaging the chip in the long haul?

My goal of this project is getting a SiS 486 motherboard working well with a Voodoo3 3000 PCI, an AMD X5-160ADZ, and 128 MB of fully cached RAM. So far, only the cache timings are holding things up. I'm also not sure how well the Voodoo3 tolerates a 40 MHz PCI bus.

Quick comment, after flashing to BIOS v1.6 (4DPSV160.zip), the motherboard would no longer turn on. I had to use a stand-alone flash programmer to flash back to v1.72.

40mhz shouldn't be to much of a problem on the pci clock. I got a v3 2k pci (sgram) humming along with the pci clock at 41.5mhz.

I spy a stack of NES and Dreamcast in one of those pics heh :p

A couple of updates on this board:

* Clearing the CMOS data takes several days. I've never seen such high data remanence on any other board, ever. Unfortunately this makes trying out different BIOS versions and configuration changes a huge hassle, so I have stopped messing around with it for the time being. Not a big deal anyway, it does exactly what I want it to as it is. 😀

* The parallel port header works fine without rewiring (unlike the serial ports), tried it out of necessity (needed an old parallel port to make a Commodore X1541 cable work) with great success.

* Found some 16 MB FPM sticks a while ago. I like my installed RAM sizes in powers of 2 so I decided to use them to bump the RAM on this board up to 32 MB. Unfortunately the board only sees half of their capacity no matter how I mix 'n match them, so back to 24 MB it is. Each 16 MB stick was tested on a fake cache PCChips board using Memtest86 so I assume they're good, this board just doesn't like those sticks in particular. 🙁

* The hot-glued CPU cooler eventually fell down and I had to hot-glue it again to the processor. 🤣 Still haven't found a suitable one. 🙁

RacoonRider wrote:

Cool boots, man!

It's been a while so at first I thought you were talking about boot screens or something, but then I remembered that my New Rock Reactor 272s can be seen in one of the pictures (badass boots indeed!). My bedroom floor is usually a huge mess of random electronic stuff, it was well past 3 AM when I was finally ready to test the board so I wasn't keen on the idea of cleaning up just for a silly picture! 🤣

simbin wrote:

I spy a stack of NES and Dreamcast in one of those pics heh :p

Indeed! My main hobby is retro gaming so I'm chock-full of these things! I'm particularly obsessed about Famicom/NES clones. 😁

133MHz wrote:

* The parallel port header works fine without rewiring (unlike the serial ports), tried it out of necessity (needed an old parallel port to make a Commodore X1541 cable work) with great success.

I also was unable to use the serial ports with a standard DB9 cable. I tried a serial mouse, as opposed to the usual PS/2 mouse, with the cable connected both directions and on both ports, but the serial mouse was not detected in Windows 98. I figured the serial header must be wired in a non-conventional way, but the manual did not disclose the pin assignments. Were you able to determine the pin-outs of the serial headers? This is on my list of things to (eventually) do.

There are two types of DB9 cables. One that has starts counting the pins on the first row and then goes to the second, and the other that counts one on the first, one on the second, etc. If I remember well the pre-pentium stuff uses one type, and the later hardware uses the other. It should be easy to check which one it is by removing the plastic housing on the back of the DB9 port.

Markk wrote:

There are two types of DB9 cables. One that has starts counting the pins on the first row and then goes to the second, and the other that counts one on the first, one on the second, etc. If I remember well the pre-pentium stuff uses one type, and the later hardware uses the other. It should be easy to check which one it is by removing the plastic housing on the back of the DB9 port.

Sorry, I should have mentioned that I tried what I thought was both types of serial cables. At least that is what I had marked the cable as, "486 type". Maybe I should recheck the cable to ensure that my label was not misplaced on the wrong cable.

It might be obvious, but maybe it's worth checking if there's a disable serial port option in BIOS.

Markk wrote:

It might be obvious, but maybe it's worth checking if there's a disable serial port option in BIOS.

The option does exist and it is listed as enabled for both serial ports.

feipoa wrote:

Were you able to determine the pin-outs of the serial headers? This is on my list of things to (eventually) do.

I did, it goes like this:

1IDC   DE9
2 1     1
3 2     6
4 3     2
5 4     7
6 5     3
7 6     8
8 7     4
9 8     9
10 9     5
11 10    x