Gigabyte Ga-880ga-Ud3h MOBO CMOS Reset works, but BIOS doesn't POST

[kanidrive]

I don't usually post my issues to Tom's Hardware, but this one has me a little stumped, and I don't see anyone else having the issues I am with this Gigabyte-Ga-880ga-Ud3h (revision 2.2) with an AMD x6 II CPU and 12800 DDR3 RAM.

The rig I built about 4-5 years ago for a client had a power supply go out (I suspect from being kept in a dusty warehouse). I replaced the power supply, everything was good again. During the upgrade to Windows 10, It suddenly started having issues rebooting. I played around with the BIOS settings a bit, but no luck fixing it. I decided to add a few more cables to the semi-modular 900-Watt Bronze Cert-ed power supply in case it was sucking too much power from one power rail. I go to turn it back on, hoping the the BIOS will POST now, but then....

..........

Nothing. At All. The fans are running, the power is good, but I'm not even getting the POST failure. Just... darkness on the monitor.

So, I've had this sort of thing happen before in my 17 years of custom building PCs, and for reasons still often unbeknownst to me, a good ole fashioned CMOS Reset can have magical effects on a stubborn mobo when you're all out of other options.

Resetting CMOS... Boom. I get the a POST message with options to reset BIOS to Default Profile, or "Last Know Good". Awesome. "Last Known Good" it is.

Blackness... again... I hit the Reset button after a few minutes. Blackness... again... Hm.

I reset the CMOS. POST / BIOS Reset options are back. Let's try "Default" this time. Wait... ... Blackness. Again. Hit the Reset Button again. Blackness Again.

I pull all 4 modules of RAM out, disconnect any fans other than the CPU, disconnect the DVD-ROM drive (I've seen these components go bad and cause issues like these before, so just being safe). There are also no expansion cards in this mobo at all anymore, so all PCI/PCIe slots are empty as I'm using the onboard graphics.

I reset the CMOS again. Blackness. No options at all this time. No beeps either. Nothing. Nada.

This one has me a bit puzzled. This machine was working well for about half a decade, all I did was swap out the Power Supply with an even bigger one than I originally had in it, and... ?????

So I DO have power,
I DO know the Video is working because I see the options after a CMOS reset.
ALL non-essential devices, cables and connectors are disconnected (Just RAM, CPU, keyboard w/o mouse & a VGA cable)

Weirdness.

If any of you can give me some new ideas to test out, I'd appreciate it. I haven't been this stumped in awhile.

 

[kanidrive]

Ha, so... a little progress, maybe? Or at least confirmation that the MOBO isn't toast...
I added back the RAM Modules, and did another CMOS reset, and during the choices to set the BIOS to Default or Last Known Good, there is an option to "continue with boot" after you toggle the options cursor/selector highlighter with the up/down arrows on the keyboard.

Doing this, I was able to successfully boot into Windows 7.

BUT...

Because I reset the BIOS settings, there were automatic uninstalled driver detection, probably because its not set to the original AHCI config I installed the OS with to begin with.

Upon reboot, I go into the BIOS to try to set everything back up the way it was before....

annnnnnd.... FREEZE.

I lose the ability to make any changes in the BIOS screen, I hit the Reset button...

Then we are back to were we were before.

Super. Duper. Weird.

 

[Calvin7]

Try replacing the CR2032 battery. They are good for 5 or 6 years, maybe longer.
http://www.batterybob.com/products/252-Maxell-CR2032-5-Batteries/?gclid=CMu8s6msm8wCFcskhgodVFILkA

 

[kanidrive]

ha, funny you mention that, because I thought I had solved the issue by doing exactly that. I had a spare lying around and popped it in, and I was able to reboot and get the "POST ERRORS" message and options menu again without resetting the CMOS via jumper pins.

But as soon as I started to celebrate by rebooting again,

Blackness. Again.

But excellent suggestion @Calvin7. I really thought that was it, but apparently not.

I've tried doing a CR2032 swap on stubborn boards in the past, but never actually had a situation where that was the problem and solution. I thought this might be the first. :/ I wish it were.

I'm running out of things to try, it's just super bizarre that it seems "after a certain amount of time" the system just craps out completely, and ONLY "after a CMOS reset" + F1 option to continue to boot, will it do ANYTHING. I just tried to download and re-flash the latest (and porbably last ever updated) F7 BIOS from the Gigabyte site: http://www.gigabyte.com/products/product-page.aspx?pid=3628#bios

But in the middle of trying, the system just totally shut down and powered off abruptly. Things keep getting more and more weird.

In typing this though, it reminds of some kind of a heat control issue. I don't smell anything cooking, and the fans are all running though... hm....

 

[kanidrive]

well, since I'm out of ideas for the obvious stuff, I started poking around a little and sure enough the heat sink on the CPU and the Northbridge were EXTREMELY hot to the touch after I let the system crash again after "jiggling the handle", so to speak, and getting everything running again via the steps I already described.

Odd though, because the fan is spinning very well, and the heatsinks look like they are in perfect position.

The CPU heatsink is one of those "sealed watercooler with a fan over the CPU and copper pipes" types that were around before the "brick" CPU coolers with the longer, fatter hoses came about. It's a nice size, lots of fins, but I do see quite a bit of caked on dust through the fan blades. I am going to pop it off to invest that more. This had all the symptoms of a heat-related issue, and now I found the course. I just have to figure out what's causing the heat. I might try to clean it out a bit and see if that helps.

 

[kanidrive]

So...

I took out the heatsink/fan unit, poked at the caked in dust and crud under the blades with a chopstick to get it loose, and stuck the vacuum cleaner hand hose over and all around the blades until it was spinning so fast it started shaving the edges of the vacuum cleaner. I got about 90% of it out.

I also noticed I has some crappy white heat sink grease on there, not sure if it was was came with the heatsink under the tape, or I added it on there myself from the tube that came with it when I ordered it, ... but either way, I scraped it off and added what Arctic Silver I had left in my stash from my last build.

When I put the heatsink + fan combo unit back in place, I did another CMOS reset before I powered up the machine. It didn't come on right away, so I had to fiddle with things again at first, but now its been running for about 3 hours without shutting down, and I've been able to reboot successfully without needing to reset the CMOS ever since.

WOW.

Not sure if I've ever had a dirty heatsink fan be the cause of such weird BIOS / CMOS behavior before. I wonder if I didn't re-ground something somewhere in the back plate maybe?

I am dumbfounded by how cleaning the fan and re-greasing the heatsink has seemed to have solved the problems I was having.

I've seen CPUs get hot and shutdown the system before, but never cause POST errors. And I've never seen a heatsink fan spinning that fast and that well not be enough to do the job.

I guess I learned a new "gotcha" today. Typically the BIOS message will complain about CPU temps nowadays, but maybe this was a board that was "before that time".

But I seems like I got it worked out. I will see if it gets through the Windows 10 Update this time or not to be sure though.

 

[kanidrive]

After some troubles with getting the Windows 10 Update to work, and about 30 reboots... There haven't been any other issues after cleaning the heatsink fan and putting some new thermal grease between it and the CPU. Wow. Weird, but true.

And Awesome.

😉

Problem Solved.