Basic Hardware troubleshooting
Sometimes machines stop working, or you install a new piece of hardware and just can’t boot again. That’s when you have to get back to some basics to troubleshoot what the problem is. Yesterday I saw a machine that wasn’t booting. It would display the splash screen and then power down. They had a new video card on a previous diagnosis using the indicator lights on the back (Dell). The add on card seemed to work, but the system was just not booting. Someone had suggested to them that it was a power problem.
This made sense in a way, old 250 watt power supply, trying to power up the system that it new, plus an addon 3d accelerated video card – maybe that was too much for it. (Original video was onboard). Maybe the power supply was going bad too…. So, I tested a new power supply, same before – splash screen and then power down. At this point I had already pulled the hard drives and cdrom’s. You know even without those the system should at least post and let you get into bios settings, or try to boot and complain about not finding an operating system.
So, I started looking at the hardware installed. The only other add-on card was an old modem. I took that out and tried, nothing… Then I thought I’d take out one of 2 memory sticks. (Total of four slots in the system.) It booted. Over the course of thirty minutes of testing it would consistently fail to boot with the second stick in it’s original “slot”. Later testing in an other slot, the system seemed to do fine.
Further background on the system is that the “no boot” problem started occasionally and then gradually got worse. Inspection of the system board found several “bulging” capacitors that I mentioned to the owner as a sign that the problem may recurr and get worse. See this article for reference. The experience of the problem makes sense with the capacitors going bad.
Anyway, the bottom line is the system was working when I left because of basic hardware troubleshooting. That means…. disconnect everything that’s not ESSENTIAL. All we need for a computer to post and try to boot are… system board, cpu, one stick of memory, video (might be onboard) and power. If one of those is suspected bad, compare against a known good substitute until everything that’s possible has been checked and by process of elimination you know where the problem is. In this case, the system board is suspect confirmed by 1)bulging/potentially leaky capacitors and 2) memory working in one slot but not another. Fortunately we had another stick of memory to test in the same slot which (again) left the system board non-booting.
With some situations you don’t have spares, but you may have other “slots” to plug a component in to (PCI slots/memory slots) when you have options, test them if it helps you determine which component is bad.
The only other bit of advice is to keep track of what you’re changing each time. If you change more than one device at a time it’s easy to get confused as to what has been proven to not work, so it’s best to just make one change at a time. If you get a combination that seems to work, test and retest to see if it’s consistent. Usually the system board/cpu is the combination that you just can’t replace that easily, so if you’ve swapped everything and are still getting inconsistent boot/no-boot situations – well it’s likely the system board.