Computer power-on issue

Help me diagnose this odd problem. I powered down my computer before a storm. Turned off the surge protector which also power all peripherals. Next day I turned on the surge protector, and the computer’s power supply kicked on, powering up the fans and LEDs. This is not normal behavior; that shouldn’t happen until I hit the power button on the computer. So I’ve got power to the case but pressing the power button does nothing at all and the computer won’t boot (or POST or anything else). I can’t even hold the power button to force it to power down.

I built the computer myself but but am not sure how to diagnose this type of problem. I am guessing that somehow or another there is a problem with the motherboard, that there is a short or something causing a “power on” signal to go to the power supply.

I have a multimeter and a rudimentary understanding of electronics. Any advice on how to troubleshoot this would be appreciated.

Is there an off/on switch on the back of the power supply? If so try flipping it a few times.

The PS On wire is usually green and connects to the motherboard with a 20 or 24-pin connector. It’s usually the largest pin on the connector. See of that pin looks OK. Just plugging and unplugging the connector may help. Same goes for the connector to the power on button.

IIRC, when you turn the thing on, the motherboard pulls “PS_ON” (usually a green wire) down to ground which signals the PSU to turn on. Then the PSU is supposed to assert “PWR_OK” (gray wire) to signal the motherboard once power is stable and it’s OK to boot. I don’t remember what it’s supposed to show on that wire during normal operation but if it’s not asserting PWR_OK then the system won’t boot.

Well, that was weird.

JerrySTL, you got me to thinking. I turned off the PS switch as you suggested, and surprisingly the case LEDs and fans stayed on. I *unplugged *the PS and they *still *stayed on. Then when I unplugged a powered USB hub from the port (connected to the motherboard, not the extended ones on the case), they turned off. So I guess parts of the machine were getting limited power through the USB port from the hub. I didn’t know it could work like that.

So now, with the USB port unplugged, I was able to power up the computer normally.

Do I have anything else to worry about with that USB hub?

(Also thanks for the PS On pin info; I was aware of that and jumped that pin to test a PS on another machine. Turns out the PS was bad and replacement solved the problem.)

It shouldn’t. If it was backfeeding/leaking voltage up to the host port, it’s broken. and it explains a few things, as that would have messed with the PSU’s stability.

I’d recycle it.

Weird indeed. You often see the LED on a network card blinking even if the computer is powered off, but not anything else. I have an external powered USB hub that’s left on even when the computer isn’t to recharge some devices. I’ve never seen it keep lights on the computer on.

Like jz78817 said, it’s time to recycle that USB hub. However there’s a possibility that a new hub may do the same thing. In that case it’s probably something on the motherboard unless you have the hub plugged into a USB card. Then maybe the card might be at fault.

One thought: Is the USB hub powered through the surge protector? If not, it might have gotten zapped in the storm.

The USB is a month old, and is plugged into the surge protector. I powered down everything before the storm, including switching off the surge protector (didn’t actually unplug it from the wall). The USB hub otherwise works just fine.

A power controller determines if and when the PSU powers on. It has numerous inputs including a front panel power button and inputs from peripheral cards.

Apparently your USB card is ordering the controller to power on a PSU.

Meanwhile, why would a defect be created if the card was not originally defective? Well a power switch only disconnects one wire. Leaving two other wires for a protector to connect a surge to earth destructively via the computer. Assuming that protector protects from that type of transient is invalid. Assuming an open switch will stop a typically destructive surge is also an invalid assumption.

Unanswered question: did a surge connect to earth destructively via that USB port and associated card slot? To answer that means describing everything connects to the computer - using a building wide perspective. How was the rest of the building connected to something that connected to that USB port?

No, because the fans and LEDs come on even if the PS is unplugged from the wall.

Still, the power controller must order the PSU to power on. Or is that power coming from something else - not from the PSU?

Its been years since I saw one. Is your PSU one with a battery inside it?

The PS is not powering on. It doesn’t have a battery. What I gather from the other posts is that the USB hub, which requires its own power supply, is leaking back voltage into the USB port on the computer, providing power enough to light the LEDs and run the cooling fans.

AIUI many systems/motherboards have the +5V USB output wired directly to the main +5V power rails. Which is in violation of the USB 2.0 specification (should be current-limted to 500 mA max) but does allow people to directly connect most 2.5" hard drives without needing an auxiliary power supply. The system fans, however, tend to connect to 12 volt power so I’m wondering how keeping voltage on the 5 volt side would affect the 12 volt…