PC only turns on sometimes, but works perfectly when it does.

[MrMartyD]

I think I have a PSU problem but am wondering if anyone has experienced anything similar, or if anyone can shed some light on this before I buy a new unit.

My PC sometimes doesn't turn on when I press the power button, but eventually when it does start it runs perfectly for as long as it's on. Yesterday it was running for 13 hours straight and not a hitch.

This morning I turned it on for 5 minutes and it worked fine, then this afternoon it was dead again. This has been going on for about 2 weeks now.

I've tried booting with just one stick of RAM, no GPU etc, but no good. I've ruled out the power switch, the motherboard is MSI and the BIOS flash button at the back also doesn't power anything up. I'm thinking that rules out the CPU as well seeing as you don't need one installed to flash the BIOS?

I've replaced the CMOS battery, vacuumed and dusted, tried different power sockets and different leads, checked all the connections etc. So it's down to the MOBO or the PSU.

I did the paperclip test on the PSU and nothing happened, so I'm guessing that's where the problem lies. The thing that's bothering me is why is it sometimes working and sometimes not, and when it does work why is it absolutely flawless whether I'm working or doing some heavy gaming? Shouldn't a PSU just die, or cause shut downs?

I'm just curious if anybody else can think of something I may have missed before I buy a new PSU, any pointers greatly appreciated!

Windows 10
MSI Gaming Pro Carbon B450
AMD 3700X
16 GB DDR4 3200 MHZ
Sapphire Pulse Vega 56
Seasonic 750W Snow Silent 80 Platinum
M2 SSD 1TB
Samsung 850 Evo SATA SSD 500 GB
SSD/HDD Hybrid 500 GB

 

[Caring1]

Did you do the paper clip test on the PSU with it disconnected from the Motherboard?
My money would be on the Motherboard being some how faulty.

 

[Ferrum Master]

Try wiggling the ATX 24 pin power connector when it happens.

 

[MrMartyD]

Did you do the paper clip test on the PSU with it disconnected from the Motherboard?
My money would be on the Motherboard being some how faulty.

Yeah I had the PSU out of the case completely with just the 24 pin connector attached and the paperclip.

 

[Caring1]

Yeah I had the PSU out of the case completely with just the 24 pin connector attached and the paperclip.

If it was connected correctly and had power, then at least the fan should have spun up.
It sounds like something may be faulty with that unit.
A cheap PSU tester can tell you if the voltages supplied are within range, if it gives any reading at all.

 

[dorsetknob]

If it was connected correctly and had power, then at least the fan should have spun up.

Most modern PSU's also Require a potential load before they fully power up
a 120 case fan is enough of a load for powering up.

 

[MrMartyD]

If it was connected correctly and had power, then at least the fan should have spun up.
It sounds like something may be faulty with that unit.
A cheap PSU tester can tell you if the voltages supplied are within range, if it gives any reading at all.

Nope, no fan, not even a twitch. Tried different leads and sockets but no joy. Good shout on the tester, just ordered one to arrive tomorrow. Ordered a new PSU as well which I can always return.

I'm just a bit miffed as to why it would work sometimes with no issues and then not at all. Either way I'll find out tomorrow. Cheers.

Most modern PSU's also Require a potential load before they fully power up
a 120 case fan is enough of a load for powering up.

Yeah I thought about this but I'm not sure how to attach a fan to the PSU, I don't have the right connectors. Is there anything else I can try?

 

[dorsetknob]

Yeah I thought about this but I'm not sure how to attach a fan to the PSU, I don't have the right connectors. Is there anything else I can try?

SATA only PSU ? just connect a hard drive/dvd drive with a disk inserted power only (you can feel/hear if it spins up)

PSU may have some Capactiors dying which could explain why it sometimes works

 

[MrMartyD]

SATA only PSU ? just connect a hard drive/dvd drive with a disk inserted power only (you can feel/hear if it spins up)

PSU may have some Capactiors dying which could explain why it sometimes works

Ok cheers I'll plug the hard drives in and see what occurs.

 

[Bill_Bright]

Verify all your power connections are tight and secure.

Since that PSU ran for 13 hours without a problem, it does not sound like the PSU to me. But it sure would not hurt to swap in a spare PSU to make sure.

Does your case have a reset button on the front panel? If it does, try pulling the two wires connected to the motherboard from the front panel Power button, and connect the two wires from the front panel Reset button it their place. Then use the Reset button as your power button and see if it works correct.

It is not uncommon for those power buttons to break, or the mount becomes loose and then the contacts when pushed become intermittent, often due to dust collecting inside.

For those who work on computers frequently, I recommend getting one of these power cable/switches. These are really handy for when you assemble the computer outside the case (breadboarding) for testing or whatever.

Most modern PSU's also Require a potential load before they fully power up

Ummm, no. Sorry but that is not correct - at least not for ATX PSUs - though I might be misunderstanding you since I am not sure what you mean by "fully" power up. But they will (if working properly) fully start without any load. I note some PSU makers, like EVGA for some of their PSUs, even throw in a little "Power ON Self Tester" jig that simply shorts the two pins needed to power on. You can also just buy them separately as seen here.

But that only tests to make sure it powers on. It does NOT test to make sure all the voltages are present and within allowed tolerances. It does not provide a variety of realist loads. And it does not test for ripple. Any power supply (including a computer PSU, car engine, battery, you name it) they all need a proper load to be properly and conclusively tested.

 

[MrMartyD]

Verify all your power connections are tight and secure.

Since that PSU ran for 13 hours without a problem, it does not sound like the PSU to me. But it sure would not hurt to swap in a spare PSU to make sure.

Does your case have a reset button on the front panel? If it does, try pulling the two wires connected to the motherboard from the front panel Power button, and connect the two wires from the front panel Reset button it their place. Then use the Reset button as your power button and see if it works correct.

Thanks for the tip. I swapped the connections and tried powering with the reset button but no dice, I can't get anything from it at all. I'll just have to wait until tomorrow when the new unit turns up, I don't have access to another one so got one on order. Hopefully it is the PSU, the idea of swapping a motherboard doesn't fill me with joy right now!

 

[Bill_Bright]

I'll just have to wait until tomorrow when the new unit turns up

Well, keep us posted.

If the 2nd supply does not resolve your problem, I would consider keeping it anyway since you have already demonstrated you are one to get your hands dirty. It is always nice to have a spare supply handy - especially since you probably have already been tagged as the neighborhood/family/friend/friend of a friend goto computer guy.

 

[Caring1]

Since that PSU ran for 13 hours without a problem, it does not sound like the PSU to me. But it sure would not hurt to swap in a spare PSU to make sure.

Deferring to your knowledge here Bill, would a faulty Capacitor cause it not to start if totally drained, yet run fine once it retains some charge?

 

[Ferrum Master]

Deferring to your knowledge here Bill, would a faulty Capacitor cause it not to start if totally drained, yet run fine once it retains some charge?

Pointing finger at +5VStby cap...

 

[MrMartyD]

Well, keep us posted.

New PSU arrived this morning, swapped it out and now it boots up first time every time.

The other one is still under warranty as it's not even 18 months old so I'll swap it and keep a spare, unfortunately I didn't have time to wait for the exchange process as I have work to do. That's the first Seasonic I've ever bought but I've been reading it's unusual for them to fail so quickly so I guess I just got unlucky.

Either way all's well that ends well. Thanks all for the advice!

 

[oldwalltree]

In the future you could go into event viewer and look at the critical errors and it will give you some codes. A lot of the time they are generic but sometimes it can be helpful and point you too the problem.

 

[Bill_Bright]

New PSU arrived this morning, swapped it out and now it boots up first time every time.

Great news! I am glad that worked.

That's the first Seasonic I've ever bought but I've been reading it's unusual for them to fail so quickly so I guess I just got unlucky.

It is rare. In fact, PSUs in general are typically very reliable. The problem is, we don't know if the PSU failed because it had a faulty component inside, or if Mother Nature (or a faulty high wattage appliance in the house) slapped it around too hard and threw an excessive surge or spike (or several surges and spikes) at it. I am a strong proponent for having all computers on a "good" UPS with AVR. Note that backup/battery power during a power outage is just a "minor" bonus feature of a "good" UPS with AVR. It is the AVR (automatic voltage regulation) that make them so valuable and important. Remember, a surge and spike protector is little more than a fancy and expensive extension cord.

would a faulty Capacitor cause it not to start if totally drained, yet run fine once it retains some charge?

I cannot predict what a "faulty" cap will do without knowing exactly how it is faulty. And even then it would be hard to predict how it (or that circuit) would behave. Back in my radio maintenance days, we used call such unpredictable behavior "FM" and it was not for "frequency modulation" but for a certain type of magic!

It is important to remember that, by design, circuits that contain such storage capacitors are designed to bleed off stored voltages (to safe levels) almost immediately once power is removed. The longer power is removed, the greater the decay until eventually (as in several seconds, not minutes) it is essentially down to 0V. This is to ensure you don't get zapped if you stick your tongue inside a PSU case even if the PSU is pulled from the computer.

And while the "time constant" (TC) value for different capacitors vary by design, the general rule of thumb is the cap will reach ~63 - 66% charge in 1TC and ~5TC to be fully charged. So my point is, it takes a few seconds (4 - 5 "time constants") for a given cap to become fully charged anyway. And that would be from a totally "cold" start - by "cold", I mean with the PSU unplugged from the wall. Remember, if simply plugged in (and master power switch set to | or "On"), that ATX PSU is running and supplying +5V to the computer already - meaning the caps would already be charged.

Clear as mud, huh?

 

[qubit]

I'm just a bit miffed as to why it would work sometimes with no issues and then not at all. Either way I'll find out tomorrow. Cheers.

There's a trigger circuit in the PSU, so if that goes faulty, this kind of intermittency can certainly happen with the PSU running fine otherwise, frustrating as it is. Shame it's on a Seasonic though as it's such a top brand.
Glad you fixed it with the new PSU. Happy gaming.

 

[HD64G]

Even a faulty cable can result in that behaviour. Has happened to me before.

 

[MrMartyD]

There's a trigger circuit in the PSU, so if that goes faulty, this kind of intermittency can certainly happen with the PSU running fine otherwise, frustrating as it is. Shame it's on a Seasonic though as it's such a top brand.
Glad you fixed it with the new PSU. Happy gaming.

I can't complain. I've been computing one way or another since the Commodore 64 and this is the first ever component that's died on me in all those years. I'm just happy it's not the motherboard!

Cheers!

 

[qubit]

I can't complain. I've been computing one way or another since the Commodore 64 and this is the first ever component that's died on me in all those years. I'm just happy it's not the motherboard!

Cheers!

About the same time for me. You’re unbelievably lucky to have only one hardware failure in all that time.

I remember an Asus mobo trashed its bios once on 2003 with an overclock too far. Totally bricked it with an RMA the only option.

 

[Mysteoa]

Is your room warmer when the PSU eventually boots or is the same room temperature?