Computer won't power on

[kalmquist]

One of my computers turned itself off, and nothing happens when I press the power button. So I tried to come up with a list of possible problems:

1. No AC wall power
2. Dead power supply
3. The motherboard circuitry that tells the power supply to turn on has died.
4. Power button is broken or not connected to the motherboard.
5. Power supply refuses to start because of a short circuit.

I unplugged the power cord from the power supply and used a multimeter to verify that there was power, ruling out possibility 1. I then plugged the cord back in and disconnected the 24 pin power connector from the motherboard. The multimeter confirmed that the power supply was producing 5 volt standby power. I then used a staple to connect PS_ON (green wire) to ground (black wire), and the power supply turned on, so I think the power supply is fine.

That leaves possibilities 3 through 5. I guess the next step is to disassemble the computer and see if the motherboard can turn on the power supply when out of the case.

In any case, I'm posting this here because if my list of possibilities is complete it might be helpful to someone else, and if it's not complete having someone point that out might be helpful to me.

 

[Ketchup]

So how long has this computer been running without issue?
Your troubleshooting steps are good. Personally, I would change the next step to:

Take everything non-essential off the board, unplug all USB connections, and see if it does anything.

Do any fans spin, even for a second, if you try this.

Some details about your hardware my also be helpful here.

 

[kalmquist]

Thanks for your reply.

I bought the system in June 2009. I've made some upgrades since, but the motherboard and power supply date back to the original system.

GA-MA790XT-UD4P (rev. 1.0) motherboard
Phenom II X6 1090T, 6 core, OC to 3.5 Ghz
2 x 4GB G.Skill DDR3 memory 1333Mhz
CORSAIR CMPSU-400CX CX Series Power Supply
EVGA GeForce 9500GT video card
ANTEC Solo Black/Silver Mid-Tower Case

If I had seen your post before I removed the motherboard from the case, I would have followed your suggestion to test a bit more first. At any rate, with the motherboard out of the case and everything removed except the CPU and memory, shorting the pins for the power button connector doesn't do anything. I tried connecting a meter to one of the molex connectors to see if the power was coming on very briefly, but didn't see anything.

One thing that I wasn't expecting, but which may be normal, is that measuring the voltage across the power button pins when they are not shorted doesn't yield a constant voltage. Instead there are pulses approximately once per second.

 

[Ketchup]

Have you tried a CMOS reset yet?

 

[kalmquist]

I hadn't thought to try clearing the C-MOS, but I've done it now and it doesn't seem to make any difference.

I've played around with things a bit more, and discovered that when the computer power supply is connected to the motherboard and I get short pulses about once every second on the 5 volt line. These didn't show up when I tried a digital meter, because the measurement speed isn't fast enough, but they can be seen with an analog meter. It seems like the board is turning the power supply on and then immediately turning it off, before the power supply has time to deliver any significant amount of power. This is without shorting the power pins that connect to the power button, so the motherboard shouldn't be trying to turn on the power supply at all.

I've thought about the possibility of stripping a bit of insulation off of the green wire and connecting it to ground to force the power supply on while connected to the motherboard, but that seems like more trouble than it's worth. The Phenom II X6 is dated enough that it makes sense to buy a new CPU (Haswell) and motherboard.

 

[Ketchup]

The upgrade is a good idea. There is something shorting you out somewhere along the line. Since you are out of the case, that short is going to be quite difficult to find IMO.

Don't forget to check with the folks over in General Hardware for upgrade advice. Even if you have something solid put together, they are a great help at finding something even better, or cheaper.

 

[kalmquist]

OK, it was the power supply after all. My guess as to what happened is that a capacitor used in 5 volt standby power circuit started leaking current. With the motherboard connected, the total draw on the 5 volt standby circuit, consisting of the power consumed by the motherboard plus the power leaking through the capacitor, exceeded the power supply limit, causing the power supply to shut down. Without the motherboard connected, the only load on the power supply was the leakage through the capacitor, which was not enough, by itself, to trigger the overload condition.

The power supply has no longer delivers 5 volts on the 5 volt standby line, even when not connected to anything. Presumably, the capacitor has continued to deteriorate, and now leaks enough current to trigger an overload condition even when the power supply is not connected to a load. (I do see a bit of voltage on the line, from the power supply repeatedly trying to start up only to detect the overload condition.)

The moral of the story is that just because a power supply works when not under load doesn't mean that the power supply is not the problem.

 

[John Connor]

May I ask what brand of PSU that was?

Edit- I see it's a CORSAIR CMPSU-400CX CX. That is a good brand. I have been lucky with Antec.

 

[Ketchup]

Glad you figured it out kalmquist!