1

There are similar questions on this site, some do not have definitive answers while in other cases it was the motherboard that was faulty or overheating issues. In my case it’s the PSU.

The purpose of this question is to take further deep insight into the way ATX Board & PSU work together.

Within 25 – 30 minutes of starting, PC just shuts off as if power button is pressed. It’s an old PC with Socket 754 board and AMD Sempron CPU.

I eliminated the following and below is the necessary information

  1. There was no overheating. Temperature never crossed 40 C (104 F)

  2. Motherboard was not faulty. This issue did not occur with another PSU. With healthy PSU, system ran continuously for 3 hours and I tested it 4 more times for that much long duration

  3. PSU in question was 450W ATX SMPS. Not from international brands like Cooler Master or Corsair or so.

  4. All HDD, DVD removed, system was just at BIOS and nothing more

  5. Once abruptly shut down, it would fail to start again for at least 15 – 20 minutes

  6. PSU fan spins well when it starts

Obviously as mentioned earlier, I have nailed it down to faulty PSU. However the same problematic PSU in paperclip test passes and using my DMM I have checked all voltages they are just good for over an hour.

What are the reasons that the system shuts down under this scenario. Could it be that the…

PSU loses PowerOK after sometime under load?

Voltages fluctuates beyond what is tolerance stated by ATX specs, under load but not when only fan is connected and no motherboard?

Anything to do with temperature within PSU?

Further, what is that integration of ATX board and ATX PSU that results into this behavior i.e. abrupt shutdown of PSU?

3
  • 2
    A PSU that "works well" with no load but is questionable with a load is IMO a PSU that is faulty and needs replacement. Passing a no-load test only indicates that the PSU is not dead-as-a-doornail. It is not an accurate indicator of the true condition of the PSU. You need to test a PSU under full load. It's like buying a used car, and accepting that it "works" since the car's engine does start. Wouldn't you take the car on a road test to be sure it's driveable?
    – sawdust
    Nov 3, 2018 at 23:22
  • 1
    The only objective task you could perform at this point is test the PSU with a full load. Otherwise you're soliciting for opinions. Vote to close.
    – sawdust
    Nov 3, 2018 at 23:31
  • @sawdust I understand your point about opinion based question. This question is like asking a medical practitioner why am I having headache. There are diverse but definite reasons for this symptom. Similarly there must be definite set of reasons for this technical behavior & one of more must be true. A subject matter expert who knows the exact integration of ATX PSU with ATX Mobo should be in a position of answering this question. As described in the question, I have already driven the car and found that it's faulty.
    – rajeev
    Nov 4, 2018 at 7:53

2 Answers 2

0

More than likely it's the motherboard that shut the PSU down. Modern motherboards monitor the voltages it is being supplied with. As long as the voltages are within a certain tolerance, everything will work. It's highly probably that the PSU has a bad capacitor or other component that begins to falter over time and no longer is supplying the proper voltage. Then the motherboard shuts down the computer.

0

~~

I find it better to ignore the "450 watt" , and paper clip test, just:

  • shut the mache down

  • take out the power supply

  • install another power supply or a new power supply

  • turn on and test under full video or program load for a few days

if the machine fails again then you know the fault is not in the power supply

~~

Not the answer you're looking for? Browse other questions tagged .