CMOS/GPNV checksum bad

[dpdevil16]

Well, I recently recapped a Gigabyte GA-7VRXP rev. 2.0 mobo with new caps courtesy of badcaps.net. The recapping seemed to go without a problem, but a couple days after doing so I woke up to find the system had locked up on a black screen and wasn't responding to any input. It wouldn't even respond to the reset button on the case (which I know normally works, btw).

Anyway, after killing power to it and starting it back up, I received a "CMOS/GPNV checksum bad" error during startup (something which I encountered just prior to finding bulging/leaking caps). Anyway, I reset the CMOS settings and proceeded to boot the system up normally and I haven't encountered any issues for the last 5 days of continuous use and operation.

I guess I'm inquiring to see if this behavior is normal within a short period of time of recapping a board, or if I should proceed to resolder the replacement caps. Like I said, it hasn't happened in quite an extended period of time so I don't know what to think. Should I get a replacement battery while I'm at it?

Thanks for the input,
Dan

[willawake]

Re: CMOS/GPNV checksum bad

the system had locked up on a black screen and wasn't responding to any input

i dont like that. run some memory tests in dos and also check your psu, open it up and look at the caps. could be a software/driver issue also. check to see what is in event viewer.

when you killed power did you a) hold the power button down until it shut off or b) pull the power cord?

a) is always preferable, or otherwise the psu switch. if you did b) and you lost bios settings then change battery.

i would not recommend resoldering the caps, if your computer is generally stable the job seems good. you should have seen immediately that the recapping was unsuccessful. soldering the caps is usually not an issue really, quite easy. most damage is from removal of the old caps or shoving new caps in, damaging lead ports, traces, leading to dead board/no post.

[kc8adu]

Re: CMOS/GPNV checksum bad

look for overheating issues too.
if unsure of your soldering gently rock the caps side to side towards/away from the neg stripe.any bad joints will show up as the leads move in the holes.

[dpdevil16]

Re: CMOS/GPNV checksum bad

Well, I ran memtest86 through 14 cycles of tests (7-8 straight hours) and not one error. So memory is fine.

CPU temperatures are well within limits, so I doubt overheating is an issue.

At the time of the lockup, the system was still using the default Windows video drivers. I believe it locked up when the monitor was in power-save mode (i.e. video is turned off). I've since updated the video and chipset drivers.

Event viewer doesn't seem to indicate anything out of the ordinary.

I had to pull the power cord since the power button was also not responding (not something I typically enjoy doing). If I remember correctly, that was probably the first time that power was totally cut from the system after the caps were replaced. I tried shutting it down and pulling the power cord from the PSU in order to replicate the CMOS checksum error, but not luck.

At this point it may or may not be the battery. Perhaps I'll try leaving the machine off for an extended period of time with the power cord detached and try to start it at a later time. Either way, I think I'm forced to play the waiting game. I might go back and reinspect the joints just to exhaust all possibilities.

Thanks for the quick responses, guys.

[AK0R]

Re: CMOS/GPNV checksum bad

Bet you've never replaced the battery since you got the board. Batteries are cheap. Just replace the battery and be done with it.

[dpdevil16]

Re: CMOS/GPNV checksum bad

Bet you've never replaced the battery since you got the board. Batteries are cheap. Just replace the battery and be done with it.

The board's not that old though. And the cost of a replacement is hardly a consideration, believe me. Besides, who's to say it's the battery? Those types of batteries rareky go bad, especially in a short time frame. I've had systems that are years upon years old and they've never needed to have the battery.

I'd rather not go chasing possibilities before doing more investigating and say with confidence that the battery is an issue.

[Rainbow]

Re: CMOS/GPNV checksum bad

If you have voltmeter, check the battery voltage. If it's under, say 2.5V, replace. I've seen boards that behave VERY strange without a battery (or with a bad one) - some don't work at all and my PCPartner TX820DS just hangs randomly during POST.

[MD Willington]

Re: CMOS/GPNV checksum bad

You could also check the +5VSB on the PSU...if that is signal is bad, the batteries will not last long.

[willawake]

Re: CMOS/GPNV checksum bad

if that is signal is bad, the batteries will not last long.

can you please expand on that interesting info

[Rainbow]

Re: CMOS/GPNV checksum bad

+5VSB is used to power the CMOS and clock when available. But even when the +5VSVB is not available, the battery should last 5 or 10 years or even more - provided that the board is well designed.

[arneson]

Re: CMOS/GPNV checksum bad

My work involves replacing memory batteries almost every day, usually in devices that are about five years old.
Many are Ni/Cad and lots are Lithium, very often with acid eating the board below.
The systems that last the longest are used everyday especially with Ni/cad.
The systems with dead memory have been safely stored or not used often.

[Rainbow]

Re: CMOS/GPNV checksum bad

The NiCd batteries on older boards were a bad idea. The lithium ones are much better - cheaper, no need to recharge, user replaceable, they don't leak and damage the PCB. I've replaced dead NiCd batteries on some old boards with a battery holder for standard CR2032 - it requires some circuit modification to disable charging.

[arneson]

Re: CMOS/GPNV checksum bad

I've done that also!
I like the buttons wrapped in plastic.

[MD Willington]

Re: CMOS/GPNV checksum bad

+5VSB is used to power the CMOS and clock when available. But even when the +5VSVB is not available, the battery should last 5 or 10 years or even more - provided that the board is well designed.

That is true, but plenty of consumer grade mainboards are NOT well designed...

[Rainbow]

Re: CMOS/GPNV checksum bad

Yes, but at least some are :-) I've got 1996 A-Trend ATC-1000 board (Socket 7, i430FX chipset) and the battery is still fine (measured 3.07V) - and that's AT system, no standby power.

[arneson]

Re: CMOS/GPNV checksum bad

The early MAC's were notorious for the lost memory, dead battery.
It was an expensive replacement for a routine repair and an inexpensive battery.

[AK0R]

Re: CMOS/GPNV checksum bad

The early MAC's were notorious for the lost memory, dead battery.
It was an expensive replacement for a routine repair and an inexpensive battery.

I inherited a 5-year old Mac (from my parents) that wouldn't remember dates, times, settings, etc. I asked them if they tried replacing the battery. They said they took it to a CompUSA store, where the local tech gave it his best, and decided that it was unrepairable (without actually specifying what was wrong). So I opened it up, took out the battery, checked it (dead as a doornail), put in a new one, and it's worked just fine ever since. Well, except for the Internet connection (slow as hell, not very stable), but that's another story, and not related to the battery. My 1986 Mac 512k still works just fine, and has only had the battery replaced once.

[MD Willington]

Re: CMOS/GPNV checksum bad

ECS K7S5A / PCChips M830...lost cmos issue, was never resolved, there were some fixes that worked...millions of those boards are out there...under various names.