| Why do you know an original supply was unreliable? For example, what determines when a PSU powers on? A power controller. This controller takes numerous inputs (including the front panel push button) to determine when power goes on, when the CPU is permitted to execute, when to trigger a safety lockout, and both if or when power must be shut off.|
Neither heat, virus, or surges are typical reasons for PSU failures. For example, heat is a diagnostic tool. If a PSU has defective parts and still works in a 70 degree F rooom, then defective parts can be identified by operating it in a 100 degree F room (also a normal temperature for consumer electronics). If that PSU fails at 100 degrees, then it was defective. And probably will fail months or a year later in a 70 degree room. Actual reason for failure is a manufacturing defect.
Manufacturing defects are the most common reason for PSU failues. We all saw a perfect example some years back when counterfeit electrolyete was used in electrolytic capacitors. This manufacturing defect resulted in major electronics failures years later. Another example of how manufacturing defects cause failures. Most failures are due to manufacturing defects.
Due to protection routinely inside all appliances, most all surges are already made irrrelevant. Any protector adjacent to that supply can only do what the supply already does better. And in some cases can even compromise internal and robust PSU protection. Furthermore, potentially destructive surges occur maybe once every seven years. Surges are not hourly or daily as so many others have been told to believe. The rare and destructive surge is averts by something completely different, located elsewhere, and costs about $1 per protected appliance. What else was damaged by a mythical surge?
Back to your problem. Is a power controller cutting off power? Does a PSU have a defect? Or has some other part in the 'power system' gone defective. Did others ignore other power system components including the power controller? Well, a minute of labor using some requested instructions and a digital meter means a next reply can answer your oroginal question accurately, without 'it might be' expressions, with reasons for why that unacceptable failure occured, and stating what part specifically need be replaced. Numbers for that meter empower the few who really know this stuff.
Your second option is to do what you probably did previously. Just replace good parts on speculation until something works. This second option, called shotgunning, is why some see a same failure maybe a year later. In shotgunning, nobody can say what is a best part to replace. Your guess is just as good as others. Just start buying new parts.
Your choice. Two options are provided. The answer to your original question: most all failures are due to manufacturing defects.