BIOS settings resetting when power disconnected

Plimogz · 06-11-2012, 07:25 PM

I've managed to lay hands on the core of my old s939 setup from way back (dfi NF4 Ultra-D, X2 3800 & 4x512MB of OCZ 533MHz TCCD,

)

The system was pulled about a year ago on account of seemingly unexplainable BSODs and following tests and hardware swaps was deemed a lost cause (or at least not worth the tech time involved)

...so, I'm trying to get the old thing up and running stable again and have come up against a problem with CMOS not saving settings. So I started by putting in a fresh battery, but settings are still resetting to defaults whenever I cut the power off at the PSU switch.

And it was at which point that into I decided to query my fellow AT forumers...

'lil help?

SpeedTester · 06-12-2012, 04:42 AM

Not that it has anything to do with the bios resetting but how do the caps look around the cpu area? They look like mine batteries and should be shiny without any bulging at the tops.

Plimogz · 06-12-2012, 12:00 PM

None of the electrolytic caps are bulging and only a single one shows a small spot of what might be corrosion from a leak. And an examination of as much of the undersides as i can see doesn't reveal anything obviously wrong either

With that said, I am in fact considering recapping the board if the replacement capacitors don't run me much over 20$. Though at a glance i estimate that it would most likely be somewhat more than that on account of the number of caps on the board.

What i think I need to find is a reproducible way to BSOD which would help isolate what's happening, maybe.

Though I've never had this sort of issue with CMOS not saving before, even on boards with distinctly bulging or leaking caps.

MacGyverSG1 · 06-16-2012, 04:24 PM

I have two of those MBs. I had a problem with the date and time not keeping current (and BIOS retaining settings) when the computer was off. Replacing the battery fixed the problem.

I would make sure the battery is good and that it is seated properly. Did you check the clear CMOS jumper?

javier_machuk · 06-16-2012, 04:52 PM

I have the same problem with my current machine, an asus p8z68-v mobo with an i5 processor... i couldn't find a solution, so now it is always connected to an UPS so i don't loose the cmos settings in case of a power loss

Magic Carpet · 06-16-2012, 05:25 PM

Quote:

Originally Posted by MacGyverSG1

I would make sure the battery is good and that it is seated properly. Did you check the clear CMOS jumper?

This.

Klavshc · 06-16-2012, 06:18 PM

When the machine isn't powered externally it is the onboard battery that keeps the BIOS settings, time, etc.

Your machine most likely needs a new battery (the small round one that also goes in wristwatches).

Pull out the existing one, write down ALL the information on it and make sure to get the exact same one or weird stuff could keep happening down the line.

Do throw us a post when you get it fixed

Plimogz · 06-16-2012, 07:08 PM

I've tried using a different battery taken from another computer that has no issues with loosing BIOS settings when i pull the plug. No improvement.

I've also made sure that the CMOS jumper is in the proper position.

For the time being I'm not disconnecting the power and also using the BIOS save slots to quickly restore settings whenever I can't get around cutting the power.

Neurotic_X · 06-21-2012, 04:09 AM

Back up data or create a mirror image.. and do a fresh install? I know it's an arduous task and time consuming but it's a shot. I did the same last night and now I'm running free and clean.. no more BSOD's, no crazy bat-sh*t unexplain-ables..

dave_the_nerd · 06-21-2012, 11:36 AM

Is the CMOS clear jumper in the on position or missing?

Plimogz · 06-21-2012, 02:57 PM

The jumper is on pins 1&2 -- which, after looking up the owner's manual online, looks to be the proper position to conserve settings. Do you think I should try to disconnect power with the jumper completely removed to see if I have any better luck? My belief has always been that the jumper closes the battery circuit when in the "save" position and likely shorts the CMOS pin to ground while interrupting battery power when in the "clear" position...

As for OS, I've already installed a fresh XP for CPU testing purposes. Given that BIOS settings don't reset so long as the system is plugged in, I've ignored the failure to save settings for now.

Testing at stock CPU speed with RAM @ 200MHz and SPD timings. I've gone through 30 loops of Memtest86+ on one 2x512 kit with no errors, and the second kit threw up 2 little test#5 errors around the 15th pass of 30. Funny thing is, I remember hearing my UPS (which isn't supplying the 939 system I'm troubleshooting) give off a couple of beeps in the middle of the night, as it does whenever the grid voltage drops out of acceptable range... So my thinking has been that, likely as not, the voltage dropped way out of spec for a split second and may have caused the RAM errors...

Also, I've so far gotten P95 errors, always on the second core's thread; one after 10 hours of small FFTs, another after 27 hours of blend.

But all of this is neither here nor there... If I manage to settle the BIOS issues I'll be that much closer to either a)knowing what the matter with the machine is or b)actually resolving some of the issues annoying me, without necessarily addressing an underlying root cause.

unokitty · 06-21-2012, 07:39 PM

If you haven't already, I would try a new rather than a used battery.

Best of luck,
Uno

Plimogz · 01-28-2013, 08:33 PM

Well, here I am bumping my own thread because some free time has me fooling around with the old kit once again.

And I'm quite perplexed to find that the issue with the CMOS not saving my settings has apparently disappeared in the six months that this thing has been lying in a drawer. Since I pulled it out a week ago, not once has it failed to keep settings intact. And this is with abundant rebooting and also a couple of day-long breaks, unpowered and unplugged. Go figure.

Now I can get back to trying to see why this box was pulled for blue-screening in the first place.

I'm pretty much down to either user error (i.e functioning RAM settings were lost and it proved too time-intensive to tweak the board into real stability again -- which I could believe, given the hell of a time I'm having getting an OS to install on the thing) or some real hardware issue.

Bad caps are still prominent on the table for the time being. CPU seems the least likely. And I suppose the DDR could have gone stale, but with 4 sticks to play with, you'd think I stand a good chance of finding at least one solid slab to get me going...

Damn this thing has a lot of memory settings...