I gave the OC'ing facility a quick go. I'm a bit puzzled as to how it works, because the BIOS set up area mentions "Change CPU's clock & voltage", yet as far as I can see, there are no options to change the CPU's voltage. I successfully got Quake running @ 150 FSB, but Windows 98 wasn't so keen, and went mad before its desktop could appear.

I tried 153 FSB, but DOS wouldn't boot up. These attempts were being made with the 1.4 Tualatin chip. I don't think I'm pushing it too hard, but I can't see what the mobo is doing with the extra voltage needed for these overclocks. Is it just adding a bit more, depending on what FSB I enter for the overclock attempt, I wonder?

Actually, I'm happier with a 150 FSB overclock on an i440BX mobo. Ironically, as you know this chipset's supported FSB is just 100, whereas VIA's is 133.

If you mod a slocket to support a tualatin, will it still work with regular coppermines?

retro games 100 wrote:

I just got the Soyo SY-7VBA133U mobo, and it works! Even though the downloaded manual says that it only supports up to a 1.2 GHz Tualatin, I put in a 1.4 hoping that it had a later BIOS and would support it, and it did! Now I've got a convenient way to test any PIII CPUs in this socket 370 board.

It's my first socket 370 board, and I'll find it much easier to test CPUs in it, rather than use a Powerleap adapter. It's awkward to unplug those things, and surprisingly awkward to remove the CPUs from them. Thanks everybody for the great retro information! 😀

BTW, the board has crazy overclocking potential. The BIOS allows you to go from 66 to 255 MHz for the FSB! That's using 1 MHz increment values. I'll try a few mad experiments tonight!

Congratulations! This is one of the best VIA694 boards IMO. I've set mine up with a Geforce 256, Voodoo2 SLI, SB Live!, NEC USB 2.0 card and 3C905TX network card and it has absolutely no IRQ conflicts, pci bus slowdowns or whatever other shittyness that plagues lesser boards. It really is just plug and play with this board.

Overclocking wise, it's not as good as my Aopen AX33 was, with a P3 1ghz, I can get 138mhz FSB, and using a Tualeron 1.2Ghz, I can get 112mhz stable. All done using some really good PC133 modules which I've tested stable at CAS 2 150mhz.

But it's a nice looking, well laid out board with a nice gold pcb and silver caps which should look good in any retro system. Enjoy!

It's a good board. I like the look of its caps. They look built to last! I think this board would be even better if the BIOS had an option to adjust the CPU's voltage. I wonder why it's missing?

For some strange reason, wire modding only goes up to a max 1.65v with the Tualatin, and no increase for Coppermines. So I think it has some overvoltage protect feature incorporated.

Update: Fiddling with the Soyo 7VBA133U last night, I found that it is indeed an uber overclocking monster! The mobo resets the PCI divider at 133, 166 and 200mhz and so on (i.e. PCI dividers of 4,5, 6 etc).

Not having seen any 370 mobo able to go over 150mhz before, the possibility of having PCI bus divider settings of 5 and over in any s370 mobo completely slipped my mind- hence my earlier inability to exceed 152mhz bus speeds.

On a whim, I installed my golden cD0 933mhz coppermine (7x multi) and set the bus to 166mhz, which meant a pci bus speed of 33mhz (using the 5x divider). Success! System posted at 1162mhz. So on I went, up to 190mhz (38mhz pci) resulting in POSTing at a scrotum burning 1330mhz. 200mhz resulted in no POST- so there's the absolute limit there.

Next, this mobo will be the basis for my Ultimate, ultimate Coppermine rig- one of 3 rigs I have planned as part of the MHz-war generation- the first rigs to exceed 1ghz, 2ghz and 3 ghz (all stable), before all that jazz became irrelevant with modern multicore systems.

Actually, the last 3ghz + machine has already been built, but it exceeded its design spec by running at 4ghz. I love old cheap hardware! 😁

Congratulations! 😀 This is really great news! I am going to dig out this mobo today, and mess about with it again. I wonder if I can set the bus to 166, with a Tualatin 1.4 GHz? The CPU clock speed will be over 1.7 GHz, but it's worth a go!

Go for it! This mobo makes me wish I HAD a Tualatin 1.4. 190mhz x 10.5 = 1995mhz Woohoo! I mean, potentially of course 😊..

Just before I dig out this mobo and mess about with it again, can I just double-check something please? Am I right in thinking there's no CPU voltage setting? If that is true, do you think this board increases the CPU voltage automatically, depending on what value you set the FSB to?

The thing is, I am used to setting a CPU voltage jumper on the mobo, or a BIOS set up area value, depending on how much voltage I think a CPU needs, in order to successfully achieve an overclock. Thanks a lot.

Your right in that there's no voltage selection available in the bios, nor any jumpers. Very odd and quite stupid really for an overclocking board.

However (facepalm), I forgot to check if the mobo automatically ramps the voltages up with the corresponding increase in fsb. I know that it doesn't happen at any point up to 150mhz fsb though.

In any case, using a tually you wouldn't be so badly affected, as the SOYO's voltage protection feature only kicks in past 1.65v, which should be plenty for a Tually. You can use the wire mod trick as I have before to set this. Again, any and all wire tricks I know of to exceed 1.65v don't work with this board. Doing this will cause the mobo to either default to 1.45v or 1.65v.

I though, am a bit outta luck , as in my experience Coppermines require a fat overvolt to 2.05v to overclock really high. So I'm crossing my fingers that my cD0 will run at least 166mhz fsb at 1.75v.

Bloody hell! I put in a tually 1.4 server CPU, and set the FSB to 166, and it boots! I had to loosen up the BIOS timings a bit. I messed about in DOS for a while, and Quake timedemo works. I get 156.4 FPS, full screen with a 2MB PCI Virge 325 card. The next tests are:

Tighten up some of the BIOS timings
Try an AGP card
See if Windows 98 behaves normally
Once I have discovered the best settings, run some more benchies...

Thanks a lot for the heads up on this one. 😀 It's fun to try out new things with old hardware.

I've done some more tests with this mobo, and I need some advice please! Everything works in DOS OK. That's with the Tually @ 166 FSB. I then power off the machine, and wait a couple of minutes. If I switch on power again, I get no POST, and the mobo reboots itself automatically, and sets the FSB back to normal levels again.

But then I had a look inside the BIOS PC health area. If I set the FSB to 133 or 166, the VCore is only 1.47V. No wonder the mobo is not cold booting properly! Please can you tell me how I can force 1.65V through the Tually 1.4 GHz server CPU? I need to understand about this wire mode trick! Thanks a lot!

sgt76 wrote:

Update: Fiddling with the Soyo 7VBA133U last night, I found that it is indeed an uber overclocking monster! The mobo resets the PCI divider at 133, 166 and 200mhz and so on (i.e. PCI dividers of 4,5, 6 etc).

Awesome! Are there any other 694T boards that support this feature? I'd assume yes, since normally it's the chipset that has to actually support this feature. BUT the board the chipset is soldered onto has to have support for this as well.

retro games 100 wrote:

Please can you tell me how I can force 1.65V through the Tually 1.4 GHz server CPU? I need to understand about this wire mode trick! Thanks a lot!

The best Tualatin mod site for your reference sir:
http://home.arcor.de/frogge/pepper/tua/tuareport1.html

Tetrium wrote:

Awesome! Are there any other 694T boards that support this feature? I'd assume yes, since normally it's the chipset that has to actually support this feature. BUT the board the chipset is soldered onto has to have support for this as well.

Honestly, I have no idea. Tualatin chips and boards are very rare in these parts. This is my first Tually capable board ever.

Thanks very much! I have one question, please. On that webpage you mention above, please can you CTRL-F (find) this 4 word phrase: As a last point

In that small paragraph, the author says: "... They [the purple pins] either need to be connected to Vss or interrupted ...". If I want 1.65V to go to the CPU, I understand that CPU socket hole "Vid2" needs to be connected/bridged to CPU socket hole "Vss".

But can I also do this another way - can I simply remove the Vid2 pin on the CPU itself? Would this mean that Vid2 has been successfully interrupted, and therefore will provide 1.65V to the CPU? Thanks a lot for any advice! 😀

sgt76 wrote:

Tetrium wrote:

Awesome! Are there any other 694T boards that support this feature? I'd assume yes, since normally it's the chipset that has to actually support this feature. BUT the board the chipset is soldered onto has to have support for this as well.

Honestly, I have no idea. Tualatin chips and boards are very rare in these parts. This is my first Tually capable board ever.

Tualatin chips and especially the boards were never common here in The Netherlands either. My guess is that there were simply never that many made and on top of that, it was the start of the capacitor plague. The Tualatin boards I did find often had leaking caps and one other board (an ASUS TUSL2-C) had this annoying whining from one of it's coils. I ended up giving the board away, even though it seemed to work perfectly.
The Tualatin chips I found were usually Tualerons 1.2 and 1.3 and only once did I find a Tualatin P3 1.2. I never actually found any of the P3-s's in the wild but thankfully they are being sold cheaply on Ebay these days, NOS 😀

I've recently gotten an ECS board (yukky) and the only reason it didn't disappear on the bottom of my stash right away is because the board happens to support Tualatins! 😜
I'd even be happy if a Compaq Tualatin board happened to find it's way into my hands 🤣
Of the Tualatin boards, over here it seems that the TUSL2-C's are the most common ones.

I am going to try the "CPU socket wire trick". I will put one piece of wire inside the CPU socket holes VSS and VID2, in order to connect them. I am hoping that this action will increase the Tualatin 1.4 GHz server CPU's voltage from 1.45 to 1.65. However, there is one thing I don't understand. In these two webpages below, the VSS CPU socket hole is in a different position!

http://home.arcor.de/frogge/pepper/tua/tuareport1.html
http://www.overclockers.com/voltage-mod-for-p … the-wire-trick/

Is this because the first webpage is referring to a "slotket adapter", and the second webpage is referring to a socket on the mainboard? Should I use the information in the second webpage? I am trying to do the "CPU socket wire trick" for the Soyo socket 370 tualatin ready 7VBA133U mobo. Thanks a lot for any advice!

Success! I followed the advice on the second webpage link, listed in my post above. I put one tiny piece of wire inside the CPU sockets labelled VID2 and VSS. When I say VSS, I mean the socket labelled VSS which is *immediately next to* VID2. Not the socket labelled VSS which is at least one hole away from VID2.

I switched on power to the machine, and immediately went in to the BIOS set up area. I then looked at the PC health status area. The VCore was 1.66V - 1.68V. Groovy! I will now continue with my overclocking experiments...

Edit: Wow, Quake fullscreen time demo gives me 181.2 FPS. That's using a PCI Virge 325. I will continue with my testing, and see if an AGP card will work OK...

With the FSB set to 166, I set the BIOS RAM CL value to 2, and everything in DOS still works!

Quake timedemo full screen = 182.8
3DBench2 = 257.5
PcpBench (no lfb) = 59.8
Doom timedemo demo3 full screen = 2134/834

Tomorrow, I'll get on to testing this system in Windows 98, with SuperPi, Prime95, 3DMark tests, etc...

You sure are one busy man! 😁