Troubleshooting Motherboard Problems: 5 Real World Examples

It's been a painful few weeks here in the basement lab. To put it more precisely, it's been motherboard hell.

I've had to build (and rebuild) an unusually large array of test systems recently. Some have been built to cover interesting new products, like the Skulltrail preview and the recently released AMD Phenom X4 9850. Others were built to compare performance between platforms, such as Skulltrail versus triple SLI and Nvidia's 790i core logic versus Intel X48.

I've also built test systems recently for motherboard reviews—which haven't panned out yet due to some of the technical issues I've encountered. So fast night, as I was slugging down a shot or two of Balvenie Doublewood, I realized that my pain could be your gain.

I've had to troubleshoot an unusually large number of technical issues revolving around new and old motherboards recently, and sharing my problems with you—and the solutions I uncovered—could help you when you build your next system.

Note that one or two of these issues are still ongoing—I'm still trying to solve them. Most, however, have had fixes, though in some cases, they weren't so obvious. So let's run down the litany of pain I've encountered recently, as well as the fixes.
Memory Madness

One of the primary issues with motherboards—particularly new ones—is memory. It's always impressive how "problem memory" suddenly starts working fine when you update the BIOS.
Still, I ran into a more prosaic issue with memory on a motherboard recently. I'd built a system for a friend not long ago—a small form factor system with a single core Athlon 64, a Gigabyte of DDR, and a GeForce 6800 GT. It certainly wasn't something you'd call bleeding edge.

He called me to tell me the video card was bad. So he lugged the system over to the basement lab and we put it on the bench. Sure enough, the fans all spun up, the lights came on—but no POST (Power-On-Self-Test) messages on the display. Must be the graphics card, right?

As usual, I didn't pay attention to the primary clue: The fans kept spinning at high speed, never slowing down in the boot process. It's been my experience that a properly booting system—even one with a bad video card—will quiet down after the initial power up. But I ignored the evidence, believed my buddy and swapped in a new video card.

Powering up yielded the same result—fans spinning at full speed and no POST message on the monitor. Muttering under my breath, I opened up the case again and stared inside. I walked around the open PC, peering at the various components and noticed it.
"It" was subtle. Memory sockets all have these tabbed levers that lock the memory into place, and allow you to easily remove the DRAM module. I saw that one of the tabs was just a little bit pulled out. I pushed down on the corner of the DDR module and felt that satisfying "snap" when a module seats properly.

The system booted normally after that. So the problem wasn't the graphics card, nor was it really the motherboard. It was operator error—or in this case,

continue...............