Contact problems are cured 90% of the time by simply removing and inserting. But sometimes they are persistant.
Sometimes I have a long streak of good luck with no contact problems, and I start thinking it is a thing of the past.
Here is what happened to me recently. I fired up an old (5 years) mobo in preparation for some changes I had in mind. It worked perfectly, like it always has. So I was set to start making changes. I put a larger memory DIMM in, and a NIC card I that I used to use with it. It refused to boot ( black screen- 3 beeps.) I pulled the card and memory out and did the routine reinsertions. No boot. I restored the computer to its original configuration. No Boot. ??? A different video card. Boots fine. So the video card is all of sudden bad? The "bad" video card had two memory chips and a ROM that were in sockets, so I took them out and put them back in (reseating). The first boot did not work, but the second did. I figured I had fixed the video card contact problems. I did some torture tests under XP. Everything was fine. I then attempt a reboot, and I'm back to a black screen - three beeps, and the same no matter how many times I retry. So I do the routine with reseating the chips on the video card again, with 6 reseats to make sure the corrosion is burnished off. That cured the problem. I put in the memory and NIC, and the computer has been working for a couple weeks. This is what can happen to you with contact problems.
A contact problem can appear to fix itself. You just flex the mobo a bit and that is enough movement to get something to make contact. Or it can be temperature related. The thermal expansion can cause enough movement to intermittently get or remove contact.
The "picky" power supply characteristic could also be a contact problem (the PS connector). Cheap powers supplies have cheap everything, including cheap ATX connectors. One of the characteristics of a contact problem is once you have changed something to get it to work, you put everything back the way it was to start with, and it works too.
I had a problem like this in a power supply connector once, and the only way I knew that was it was that the plastic shell ultimately deformed and swelled. Since that pointed me to the problem, then I could feel that that one contact would get hot after say 15 minutes. If it had made really good contact, it would not have gotten hot.
Loosing the CMOS settings can be a contact problem with the battery contacts.
When I first got my ABIT KT7 mobo a couple of years ago, I did torture tests on it down in the basement sitting on a cardboard box. It would give some type of Windows error usually with an hour or two. I did everything I could think of to eliminate the problem, including reseating the CPU and PS connector multiple times. I was persuaded to buy a big PS, which were kind of expensive at the time ($90), but that had no effect. The morning I intended to RMA the mobo, I decided give it one more try. It ran all day long without any errors. I put the 250W PS back and it also worked perfectly. Since I had so much trouble, I torture tested it for ten days before I was satisfied. That mobo never gave me a poblem like that again, and it is still my main mobo. I can't prove the problem was a contact problem, but I would guess that it was, and I think it was the CPU socket. Those AMD CPUS pull massive amounts of current, and that can bring out any slight contact problem.
You might imagine that contact problems are due to inferior quality contacts, but this is not the only way it can happen. Mobos are washed with a caustic detergent after they are wave soldered to get rid of the soldering flux, and rinsing off every spec from every crevice and hole is difficult. Any corrosive material remaining can lead to contact problems. Enviro-friendly washing solutions may not clean all the solder flux off, which in itself can cause problems, because solder flux is corrosive (which is why they clean it off.) In addition mobos are shipped over the ocean, and the sea air is caustic. Electronics are put in bags and containers that may not be always sealed completely. Even in sealed bags, you can get condensation caused by a drop in temperature. They usually pack it with a little bag of calcium chloride, which absorbs mosture, to mitigate potential problems, but this is not a 100% guarantee.
