CMOS battery replaced, but mobo still losing time

[endorphin]

I purchased the Gigabyte GA-EP45-UD3P in March 2009. Recently, I noticed that the CMOS was losing time. So I thought the battery might be the issue.

But after replacing the battery with a new one, it's still losing about 10-15 mins per week. Since this has begun, I'm now beginning to get intermittent bsod. Anyone got any ideas as to what might be the issue? Is my motherboard possibly faulty?

To rule out memory, I also tested the memory for 8 hours, and nothing came up as a problem with that.

I run Win Vista 64 ultimate edition.
8gig Ram
Nvidia GTX 285
Corsair TX 850w power supply.

Any ideas as to how I can troubleshoot this would be greatly appreciated.

Cheers

 

[bilbat]

I'm pretty sure the clock is maintained by a piece of the CMOS hardware; once in a while, you'll get one that, like a crappy watch, just doesn't keep very good time. If it were a matter of the battery, you'd be likely to have other CMOS related errors, like checksum failures...

You can synchronize your computer time automatically with internet standard time server. Normally this feature is enabled by default in Windows XP and Windows Vista. If not then you can enable it now.

For Windows XP, double click on the computer clock at the bottom right corner and go to “Internet Time” tab. Under “Date and Time Properties” windows, tick the check box labeled “Automatically synchronize with an internet time server”, and you can select the alternative time server using drop down list.

For Windows Vista and Seven, left click the computer clock at the bottom right and click the blue link “Change date and time settings…”.
Under “Internet Time” tab, click on the “Change settings…” button and then tick the check box labeled “Synchronize with an internet time server”.

Now, your computer will periodically update time with internet time server but you can force your computer to immediate synchronization. For manual synchronizing click the “Update Now” button. What I haven't been easily able to find is what 'periodically' means - once a day, once a week? If it doesn't seem to work for you, post back, and I'll link some software that you can force to update at a period you can set...

 

[endorphin]

Thanks for the reply, I'll check that. Can this kind of thing be caused by other hardware issues on the mobo?

 

[bilbat]

I've looked around, and haven't found any. While I could envision writing a virus that would mess with the clock, I really don't think it's a likely 'side-effect' of anything else - for the most part, programs don't care about the real-time, they only respond to 'ticks' of the thread scheduler's time-keeping; if anything, they might, for convenience, read the time to display it for you, but never contain anything that would ever write to the time...

 

[GReaper01]

Thanks for the reply, I'll check that. Can this kind of thing be caused by other hardware issues on the mobo?

Might not be the motherboard itself. Have you seen where your other hardware is as regards to the CMOS battery, or checked to see how the battery was seated? On the Gigabyte MB I got, I had similar trouble with time losing sync. I opened it up, and saw that the battery was just about next to my video card. I readjusted that a few times just to be sure (the battery) since it appeared to be a bit off.

Also, you might want to check some of the installed software, as some of it may indeed slow the time down. Are you running the software that came with the motherboard itself? It might be part of the culprit causing your BSOD's.

Should be some helpful advice, and things to check out. Can't afford to rule out anything at this point.

 

[bilbat]

I stand by my statement - anything with a serious enough effect on CMOS Vbat to cause the clock problem would result in CMOS checksum errors; programs, as far as I can discover, are also unlikely culprits - MS themselves warn, even in the DDK docs for building ring0 drivers, against direct access to the RTC - can cause severe latency/interrupt sevice problems in the OS. The CMOS RTC (unless things have changed radically recently) operates off a phase locked loop feedback corrected oscillator that will go dead before 'losing' a tick; one thing I will do, when I next need to be on my knees 'in the box', is to examine a board to see whose circuit GB uses (likely, the same thing on all GBs - that's just how they work), and get the specs - the only possibility that I can fathom is that some (fairly rare) CMOS/RTC setups will actually deactivate the clock when the Vin gets low, to 'protect' the CMOS retention itself - but, I believe these will all write a 'dirty' value to the RTC registers to indicate that this has occured, and the clock is no longer valid...