Question Power supply outputting correct voltages but no power to motherboard

Jun 19, 2023
3
0
10
Power supply outputting correct voltages but no power to motherboard

This one has me stumped. Suddenly yesterday my pc won't turn on, I was using it earlier in the day.
I noticed the LEDs on the motherboard were dimmer than usual and flickering so I assumed it was the power supply.
I jumped the pins to power it on and tested each pin on both 24pin and 8 pin connectors for voltages and they're all correct (barring some 5v pins showing as 5.2v but that should be within spec +-5%)
So I thought it might be the motherboard, tested using another power supply and the power light comes on when the power button is pressed meaning it's powering on and the motherboard LEDs are bright again.
Put my power supply back in and now no LEDs go on but the power supply is outputting the correct voltages...

So I know it's the power supply I just don't know why it is. Has anyone got any ideas before I end up having to buy a new one?
 
recheck the 24 and 8 pins connector for any loose cable in them but look like this one is out what is the age of this psu
I don't have a reason to believe that the pins aren't connecting properly because in recent months the pc has been randomly losing power and rebooting itself so I am suspecting the PSU failed, especially since it's a good 7/8 years old I think.
It's just confusing me why all the voltages seem correct but the board won't get power, I'll have to have a look when I finish work to see if any pins have come lose and need to be readjusted when plugged in but I don't think that's the issue, I'd like the be proven wrong though
 
I assume that when you "jumped the pins to power it on" the PSU was not connected to the motherboard or any hard drives, fans, etc?

If so, you didn't have a load on the +12V and +5V rails which is not representaive of normal working. A more accurate test would have been to connect several 12V automobile (car) headlight bulbs rated at 60W to the +12V and +5V rails, to simulate a load.

With a reasonable load on the PSU, you might see excess ripple on the +12V or +5V rails (using an oscilloscope) caused by a failed electrolytic capacitor or some other component inside the PSU.

Given the PSU is 7 to 8 years old and has stopped working, chuck it in the bin before it damages your motherboard and fit a new unit.
 
  • Like
Reactions: hotaru.hino
I assume that when you "jumped the pins to power it on" the PSU was not connected to the motherboard or any hard drives, fans, etc?

If so, you didn't have a load on the +12V and +5V rails which is not representaive of normal working. A more accurate test would have been to connect several 12V automobile (car) headlight bulbs rated at 60W to the +12V and +5V rails, to simulate a load.

With a reasonable load on the PSU, you might see excess ripple on the +12V or +5V rails (using an oscilloscope) caused by a failed electrolytic capacitor or some other component inside the PSU.

Given the PSU is 7 to 8 years old and has stopped working, chuck it in the bin before it damages your motherboard and fit a new unit.
8pin was left on the board and so were the case fans but no, the 24 pin wasn't, I jumped the power on pin to ground and the psu and the case fans turn on and stays on so you're most likely right regarding it not being able to hold a real load since it won't power the board.

I'm going to buy a new one regardless I suppose, I just wanted to know why it wouldn't power the board when it showed correct voltages since another PSU will so I know its not the motherboard, it was just strange to me.

Thanks for the help
 
If the 24-pin connector wasn't plugged to the motherboard, I doubt that any significant current was flowing through the +12V CPU connector, because the motherboard wasn't switched on. The only way to be certain would be to connect a DC current clamp around the black/yellow CPU supply wires.

Remember, a typical DC multimeter set to the Volts range only displays the average voltage, not the instantaneous voltage. On a faulty PSU, there might be several Volts of ripple but you'd be hard pressed to detect it on a DC multimeter.

If you set the multimeter to AC Volts and the ripple frequency is within the range of the instrument, you might be able to measure a significant AC component on a faulty PSU, superimposed on top the DC component (which wouldn't be displayed on the AC Voltmeter). The normal ripple on a good PSU is usually in the tens or hundreds of milliVolts. See PSU reviews on Tom's for examples.

For a true reading of ripple you need an oscilloscope.