I've got the mobo stable @ 150 FSB using the 1.4 GHz tually, in Window 98. I'm using an nVidia GF3, and VIA's 4-in-1 chipset driver version 4.43. 3DMark 2001 works OK. I tried 155 FSB, but when the BIOS POST screen is about to display the line of text saying something like "Starting Windows", it doesn't show this line of text, but just hangs.

I wonder if it needs more IO volts? If true, then it's a pity that you can't increase IO voltage on this mobo.

retro games 100 wrote:

I've tried a couple of SL6BY CPUs, but I can't get DOS to boot up @ 166 FSB any more. I can't remember where I bought these SL6BY CPUs. They have a Compaq sticker on them.

This ebay advert looks familiar. What does this seller mean? "ONLY SUPPORT 815T SHIPSET MOTHERBOARD !!!!" The mobos I am testing these CPUs on is i440BX based and also this Soyo VIA Apollo mobo. They seem to work, so long as you don't try to overclock them too much!

My guess is he just doesn't want any troubles with people trying these Tualatins in boards that don't support them 😉

I used UniFlash to save this mobo's BIOS. It appears at the end of this post as an attachment. I renamed the .bin file to a .txt file. You simply need to rename it back again, to a .bin file. The relevant information is as follows -

7VBA133U-2BA2 (Bios version, top of BIOS POST screen. I think it's the latest.)
08/13/2003-694T-686B-6A6LJS2CC-00 (Bios string, bottom of BIOS POST screen)
UniFlash says: SST 39SF020(A)/5V

Judging by the BIOS date above which is 08/13/2003, I reckon a good VIA 4-in-1 service pack version to use would be 4.43, for Windows 98. Here's the link to it. Also,

Link to the mobo's .PDF manual, version 1.2
Link to the mobo's older .PDF manual, version 1.

Edit:

I found a hardware monitor application for this mobo. It's the VIA hardware monitor app, version 2.06. I don't know if it's the latest version, or not. This app seems to think my 1.4 GHz tually is running at 20C. Seems quite cool. However, the heatsink is good. Here's the link to this app.

Thanks RG100! Thoroughly comprehensive testing, reporting and summarization as usual. Unfortunately, my board seems to have problems booting - actually it seems dead as a dodo. Haven't got round to troubleshooting comprehensively yet.

No POST anymore? I'm confident the mobo still works. Please try the following:

1) Remove the mains plug from the wall
2) Put the CMOS jumper in to the "clear data" position
3) Remove the wire mod. I think it confuses the BIOS. I have now removed mine. I don't think it's 100% stable/reliable.
4) Make a hardware change. Sometimes this "wakes up" the BIOS, if it's got confused about something. So, remove your AGP VGA, and replace it with a PCI VGA.
5) Put the CMOS jumper in to the "retain data" position
6) Put the mains plug back in to the wall

I'm sure it will POST now. Good luck! 😀

You could go back to experimenting with the wire mod, but I couldn't get cold boots to work. I also had your situation where the mobo appeared to have died.

Guys , cmon... wtf - you do all that, you post @ 166 with maxed timings and you dont post Sandra 200x results? Where are the 3dmarks? Thumbs up for DOS benchmark, but we wanna see some real stuff 😜 And feel free to compare it to the KT133a XP-m @ >2ghz.

BTW, 'tis true that P3s may be saturated by SDRAM, however, its hard to find decent SDRAM modules that work with CAS 2 @150+Mhz, while its VERY EASY to find DDR modules that will accept up to 220mhz FSB with best timings, so I'll wager its easier to OC a P3 on a DDR board then on a SDRAM board (just like the case KT133a vs NF2, same athlon CPU).

Ah, but I couldn't complete my tests. You see the thing is, I couldn't get DOS stable. That's why I could not go on to do the "real stuff", and get Windows working. The mobo was not stable @ 166. Believe me, the Windows stuff was the next thing on my list of things to do.

I downloaded an era appropriate version of AwdFlash, version 8.22B (09/12/2002), and made another back up of this mobo's BIOS file. What is interesting is that I examined both saved BIOS files using a hex editor, and ran a "file difference" check on them, and they are different. My guess is that they are both OK, but you can't mix one saved BIOS file, with the another flash util. So, you can't take the back up file made by UniFlash, and then use it with the AwdFlash util. And vice versa. I think.

Link to BIOS file, made by AwdFlash version 8.22
Link to AwdFlash version 8.22
Link to UniFlash.

Why dont you try to loosen up the timings a bit, lets say completely - to cas 3 and normal instead of turbo? It might do the trick?

[quote="elfuego"]Guys , cmon... wtf - you do all that, you post @ 166 with maxed timings and you dont post Sandra 200x results? Where are the 3dmarks? Thumbs up for DOS benchmark, but we wanna see some real stuff 😜 And feel free to compare it to the KT133a XP-m @ >2ghz.[quote]

Woulda done that, but this board is a major pain inda buttocks to get stable 😜

elfuego wrote:

Why dont you try to loosen up the timings a bit, lets say completely - to cas 3 and normal instead of turbo? It might do the trick?

Yep, I tried that. I set all timings to "slowest possible".

retro games 100 wrote:

No POST anymore? I'm confident the mobo still works. Please try the following: […]
Show full quote

No POST anymore? I'm confident the mobo still works. Please try the following:

1) Remove the mains plug from the wall
2) Put the CMOS jumper in to the "clear data" position
3) Remove the wire mod. I think it confuses the BIOS. I have now removed mine. I don't think it's 100% stable/reliable.
4) Make a hardware change. Sometimes this "wakes up" the BIOS, if it's got confused about something. So, remove your AGP VGA, and replace it with a PCI VGA.
5) Put the CMOS jumper in to the "retain data" position
6) Put the mains plug back in to the wall

I'm sure it will POST now. Good luck! 😀

You could go back to experimenting with the wire mod, but I couldn't get cold boots to work. I also had your situation where the mobo appeared to have died.

You sir, are a guru. I tried your advice... and the mobo works! Many thanks o' wise one.

Great! 😀 One day I'd like to go back to that mobo, and see if I can the wire mod trick to work successfully...

I got a different chip to use on this board, a 1000A. Hopefully this will work at 133mhz FSB at least. 150mhz would be FTW.

sgt76 wrote:

I got a different chip to use on this board, a 1000A. Hopefully this will work at 133mhz FSB at least. 150mhz would be FTW.

One hint I can give you is, if you want to overclock a Tualeron, to use one of the later steppings. They "should" be better overclockers. Otoh, I don't know from top of my head what the lowest speed rated later stepping Tualeron was (meaning, I'm not sure if a later stepping 1000A was ever made).

My 1000A is an SL5ZF Phillipines chip- supposed to be a good one unlike the 1200 dud I had earlier. Tonight's preliminary testing it POST up to 185fsb with a 1.65v (actual 1.7v in bios) wire mod. I'm sure I can get 13x-140mhz stable without blinking. 150mhz would be golden and 166mhz + would be an uber 1337 "ohmyfuckinggodijustshatmyself" moment.

sgt76 wrote:

My 1000A is an SL5ZF Phillipines chip- supposed to be a good one unlike the 1200 dud I had earlier. Tonight's preliminary testing it POST up to 185fsb with a 1.65v (actual 1.7v in bios) wire mod. I'm sure I can get 13x-140mhz stable without blinking. 150mhz would be golden and 166mhz + would be an uber 1337 "ohmyfuckinggodijustshatmyself" moment.

I presume you meant 10x? These chips are all normally locked 😉

I meant 13x as in one hundred and thirty something, not 13x multi... if I had a chance of getting 166x13 = 2158mhz, then there would be no expletives in the world that I know of that could describe that! 😁

OK, update on the soyo. Some extensive test have been done this weekend, and here are the findings:

1) When using 100mhz bus speed chips, i.e. Celeron- this board doesn't give the correct pci dividers at 133mhz. So I had to pin mod my 1000a into a default (BSEL 1 mod) 133mhz chip. Then it was possible to go to 1400mhz + easily without volt mods.

2) The board doesn't like more than 1 ram slot populated. Using 2-3 x256mb modules, I couldn't get it stable. Using only 1 256mb module all is hunky dory at 1333mhz (10 x 133mhz). More than that wasn't prime 95/ memtest stable.

3) Wire mods actually decrease stability with this board.

Verdict: Finicky board this - tends to overthink itself fcuking up your overclocking with all its overvolt protection and auto fsb crap. Must say that once you figure it out though, its very stable and smooth.

I have pretty much the same problem. Right now I've got this setup:
Pentium III Coppermine 1GHz
Asus CUSL2
Nvidia GeForce 4 Ti 4200
Voodoo 2 SLI
Aureal Vortex 2 + Daughterboard
PCI LAN Card

I'd like to try the 1.4 GHz Tualatin with 512kb cache, but the motherboard doesn't support it. I'm thinking about doing the mod but I was also thinking about changing the Vortex 2 as well and replacing it with a SoundBlaster Live and a SB16 + Daughterboard, but that will require an ISA slot. I've also got a DFI PA61 board , so maybe I could get a slotket adapter but that board uses a VIA chipset and I'm kinda skeptical about it.
Any suggestions on what I should do?

No one? 😜