By using this site, you agree to our Privacy Policy and our Terms of Use. Close

I work as an IT director now, been working on PCs (and before that, programming on Atari/Commodore 8 bit) since the 80s. Over that time I've probably fixed, built, upgraded, or otherwise worked on 10-15 thousand systems.

The PSU is definitely the most common part of failure in my experience, primarily because imho it's where most of the OEMs try to save pennies by going with the absolute lowest price source. A customer might walk into best buy and see that a system with 8GB of ram is 'better' than one with 6GB of ram, but the power supply type and brand probably won't be listed, and on average the vast majority of people won't have any idea anyway. Some systems were absolutely notorious for failure such as the mid to late 00s eMachines and HPs that used the infamous 'Bestec' PSUs. In most cases when a PSU dies (say 80%+ of the time), the mainboard and other components are fine. But with those Bestecs, it was super common for them to fry the board along with various other sensitive components.

Closely following the PSU are mechanical hard drives. After that (if the system came with one, which has been rare for a long time), the video card (cheap stock fans and undersized heatsinks on the cheap video cards HP/Dell liked to use meant that heat was often an issue and the caps would bulge/fail too), which was just a little more common than the mobo failing. Then optical drives, then memory, and lastly CPUs.

So : PSU > HDD > Video Card (if any) > motherboard > optical drive (if any) > memory > CPU.

I've only seen a single failed SSD so far out of several hundred, so they seem to be perhaps as reliable as a motherboard or piece of ram.

Seasonic make great PSUs. I've also had good luck with Antec, Delta, Corsair, and Enermax. Choosing the right model for the job and keeping it clean goes a long way.