Yuku free message boards

Here's an update of my attempts to mitigate the CMOS clearing problem on my motherboard during cold startups.

Attempt #1 (replacing 1.8V supply filtering cap):
I was initially wondering if the problem was caused by a power supply sequencing issue. The 1.8V supply for the SIS735 chip has a very large capacitor on it (1500uF) so I was wondering if the 1.8V core supply was rising too slowly relative to the I/O supply (3.3V I assumed - don't have a datasheet for the SIS735 to look at). Some multi-voltage IC's care about power supply sequencing so I was wondering if this was the case here. I replaced the 1500uF capacitor with a 220uF capacitor. At first this seemed to work but after a couple of weeks the CMOS cleared :-(. I then tried a 150uF tantalum capacitor that I had sitting around in my spare parts bin. Same results as before, after a few weeks, the motherboard lost its CMOS settings.

Attempt #2 (1.7V resistor fix):
I put the 1500uF capacitor back in, returning the motherboard to stock condition. Using the resistor fix I lowered the SIS735 core voltage to 1.7 volts. This seemed to work. Then after about a month I quickly cycled the power on-off-on and lost the CMOS. The lower voltage appears to have raised the temperature at which my motherboard loses its CMOS.

Attempt #3 (1.85V resistor fix):
By putting a resistor in parallel with R274 one can raise the voltage supplied by the LD1117 regulator. This is the configuration I'm currently running and so far (almost a month) I haven't lost the CMOS (fingers crossed).

So far, everything I have tried, corroborates what Mr.Athlon said. So, if you are experiencing the CMOS problem, definitely have a look at his web page:

www.geocities.com/mrathlo...ipset.html

About my system:
K7S5A Pro (Rev 5.0) w/Athlon XP 2400 (barton) processor. 512MB DDR Ram. (The CMOS problem started showing itself after switching from 2x256MB of SDR SDRAM to 1x512MB DDR SDRAM.)