One night I normally turned off my computer, however next time when I turned it on, it didn't start at all. Fans weren't spinning, and there was nothing else(even no BIOS beeps). I managed to trace the problem to defective power supply(not enough current in PWR ON cable in main ATX connector) and replaced it. After that my PC at least starts(this time with one "OK" beep), but there is no display. When I switch to mainboard's integrated VGA, everything runs fine. The VGA was tested separately on different mainboard and run smoothly, so there's probably something wrong with my mainboard's PCI-E slot. So I tried to replace the mainboard with the brand new one from a shop and as long as there is my original VGA connected, the system won't start - fans are spinning(VGA fan too), but there is no BIOS beep, almost no HDD activity and again - no image on the monitor. If I remove the VGA, the system starts to "behave" according to expectations( 1 long, 2 short beeps = VGA failure/missing) and HDDs are spinning normally. As the new mainboard does not have the integrated VGA, i can't determine if the VGA got also somehow damaged when the original, faulty power supply was installed. However, the VGA itself did run on MSI K9N mainboard.
Original power supply: Enermax Liberty 500W
Present power supply: Seasonic X-650(W)
Original MB: Asus M3N-HT Deluxe
New MB: Gigabyte GA-M720-US3 rev.1.1
VGA: Asus EN7600GT/2DHT 256MB
And there's another thing. It's kind of strange, but for the last 4 years, the PCs(except for the notebooks or things like routers/printers...) keep failing ,with almost no exception, in this precise time of the year mostly for these various reasons: motherboard/VGA/power supply failure.
I have surge protection installed(also with the backup power supply) and no other electrical device than a computer ever failed in our house.
Any suggestions or help would be really appreciated.
Under voltages can be just as harmful as overvoltages. Your surge protector only protects you from spikes (high overvoltages). You say you have a backup power supply. I assume you mean a UPS (uninterruptable power supply). You need to get a UPS that has line conditioning or a separate line conditioner.
Nasher - you are doing shotgunning. Shotgunning routinely creates confusion.
The list of suspects is long. Too long to discuss here. Normal is for a defective part to work in some systems. Only way to identify what is and is not defective means fundamental facts and numbers. For example, only better computer manufacturers provide comprehensive hardware diagnostics for free so that you are not confused. So that facts are solid and obvious. Unfortunately, it sounds like you have a clone machine - therefore have no such diagnostics. Just another reason why confusion is easier.
One system that can cause any part to act strangely is the power system. That is more than just a power supply. Only numbers from that system (with nothing disconnected or removed) can identify that system as either 'definitively good' or 'definitively bad'. Until you take critical voltage numbers, then anything can appear intermittent or defective - create more confusion.
A digital multimeter is sold in most stores that also sell hammers. Available even in Kmart. Or from Wal-Mart for $18. In your case, a multimeter set to 20 VDC measures voltages on wires that connect the supply to motherboard. Measure any one of purple, red, orange, yellow and gray wires to three significant digits. Best done during power on and during (just before) any failure. Those with the most knowledge can provide no help if you do not provide these fundamentally simple facts.
Your system fails with an add-in video card. And works with an onboard video controller. So change to the other controller and measure again. Numbers are necessary to have an answer without doubt and without speculation.
Do not wildly speculate as your first post did. You have no reason to suspect a PCI-E bus or AC power. None, because you have no numbers or diagnostic reports. Or even information from Windows system (event) logs. And because those devices typically do not cause your symptoms. Get numbers. Then have a useful reply that IDs a suspect, or exonerates it. As said in CSI, "Follow the evidence". Don't wildly speculate. Don't shotgun.
This checks the PSU under no load conditions, so it is not completely reliable. But if it can not pass this, it is dead. Then repeat the checks with the PSU plugged into the computer to put a load on the PSU. You can carefully probe the pins from the back of the main power connector.
In your case, this should be in three steps:
1. Just PSU.
2. System using on board graphics.
3. System with video card.
You can also try to test the video card in another system.
Yellow wires should be 12 volts. Red wires: +5 volts, orange wires: +3.3 volts, blue wire : -12 volts, violet wire: 5 volts always on. Tolerances are +/- 5% except for the -12 volts which is +/- 10%. ...
You can turn on the PSU by completely disconnecting the PSU and using a paperclip or jumper wire to short the green wire to one of the neighboring black wires.
Those numbers will leave a defective system undetected. Stated previously was that those numbers must be posted here - to three significant digit. Because information not posted is found in those three digit numbers. Reasons why will be apparent after those numbers are posted.
That paper clip test reports nothing useful. A defective supply can still pass that test. An answer that is not definitive - with any doubt - is nothing more than wild speculation.
If that YouTube video is one I am thinking of, then additional damage can be created by that procedure.
That paperclip test does not result in numbers. It is taught to A+ Certified Computer Techs so that no electrical knowledge is necessary. That test causes vague and misleading conclusions - also called confusion. And it says nothing about the power supply controller.
Posted earlier, a defective supply can still boot and run a computer for months. Anyone who does not know why has not yet learned basic information.
OP will learn significantly by posting three digit numbers from each requested wire. Not just numbers; also behavior. A procedure that takes less time than the paper clip test. The procedure reports most information when conducted with nothing disconnected, swapped, or changed. What results is an answer that is definitive - no doubts - no reason to do any speculation or shotgunning.
Another myth. Undervoltage causes no damage. A benchmark indicator of insufficient electrical knowledge is that popular undervoltage myth. Undervoltages are harmful to motorized appliances - not to electronics. Sometimes we intentionally design undervoltages inside electronics to increase life expectancy.