View Full Version : Minor harware hack

March 1st, 2006, 07:56 PM
Hi all,

I have an ECS K7S5A that apparently gave the overclocker crowd a major boner some years ago. This particular board, however, has an odd problem. The BIOS hardware monitor reports the core voltage as 1.35v. It actually works fine as long as you don't try to push the bus faster than 110MHz (using an overclock friendly BIOS). The official BIOS has bus settings of 100 or 133MHz, the unofficial one goes from 100 to 150MHz in small steps.

Of course, the core voltage is NOT adjustable. But, I found a website that shows how to increase the core voltage. I figured I'd give it a shot and see what happened, and it really worked. Core voltage is now 1.79v and cpu temp at idle with WinXP Pro is 111F.

I intend to solder a resistor in place of the decade box, but I thought y'all might like to see the current setup.

Think I'll let it run like this for a while to make sure that it's stable.

March 2nd, 2006, 05:01 AM
My PC has an ECS K7S7AG GameUnion motherboard, with integrated SiS Xabre 200 graphics (and no AGP port, which means it is a dead end for graphic updates). I wonder if this board and the other you mention share any similarities. The K7S7AG has SiS 746 + SiS 963 chipset, up to DDR400 memory and a 333 MHz FSB. I've updated BIOS once, but frankly never found out if my CPU settings are accurate or not; at least BIOS reports my CPU as 2000+ like it should.

March 4th, 2006, 03:28 PM
The K7S5A has 4x AGP slot, SiS 735 chipset, up to 266MHz frontside bus and DDR, max RAM of 1GB, onboard LAN, and a few other goodies.

I was wrong about the max system clock setting on the overclock bios, top speed is 200MHz (400MHz frontside bus). Don't know if the board can really do it, having it run at 133 is good enough for me for now. Plus, the fastest ram modules I have is PC133. Looks like I'm going to have to install a system fan, it was locked up when I got home this afternoon.

I have an additional K7S5A sitting here, it has bad capacitors all over it. Won't boot at all. I'm debating whether or not it's worth it to replace the caps.

Installed the appropriate resistors yesterday. One oddity, at 133MHz it identifies my Athlon XP 1700+ as an 1800+. Weird.


March 5th, 2006, 11:40 AM
It appears like two rather different motherboards, mainly sharing the chipset manufacturer and of course CPU socket. In the BIOS setup, I can change CPU Frequency (100/133/166 IIRC), CPU Over-Clocking Freq (in steps of 1 MHz I think) and CPU/DRAM ratio (usually 1:1). Currently my board is set to 133 MHz CPU frequency, which combined with the 12.5 multiplier (not changeable?) makes 1667 MHz = XP2000+. No idea about voltages or what a different voltage would do to the system. I used to understand those things quite well with the Pentium 1 and II series, but not after that.

March 5th, 2006, 01:31 PM
Yep, sounds like two different boards, yours is probably newer.

Found a 10k ohm multi-turn pot and installed it in place of the resistors. Adjusted the Vcore to 1.71v (nominal voltage is 1.75, board was running 1.35 before the mod). I'll see how stable that setup is. If it crashes again, I suppose that I'll just let it run at 1.35v and 110MHz.

I think that I mentioned it before, but the factory BIOS had 100/133 as the only choices of bus speed.


May 4th, 2008, 05:42 PM
Here's an update to this old thread.

Got home after work today (yep, work on Sunday) and found the K7S5A locked up. A bit of testing showed that the onboard IDE controller was bad. Also, the Vcore had slipped to 1.64 volts. Bumped it up to 1.75 volts but it still didn't work right. All other parts of the board seem to be ok.

So, the old board ran for 2 years nearly continuously, turned it off only very occasionally.

I swapped out another K7S5A that I had replaced the capacitors on, it's currently running some tests.

I'm hoping that WinXP won't complain about this motherboard swap since they are the same model. Reinstalling all the crap, er, apps would be a major pain. Not to mention remembering all the setup parameters for said apps.

If I should run across an inexpensive PCI IDE Ultra-DMA controller it may be worthwhile to find out if the old board will work with it.