Power Supply 13 years old, should I be concerned?

[UltraTaco]

Technically, a used might be considered more reliable because it's proven to work, but when you hook up a brand new something, there's a good chance it'll either be doa or destroy your equipment. Beware of new PSUs because when they pop, they can take everything else with it.

 

[EarthDog]

What is this...bizarro world????

No. No. And No!!!!

 

[UltraTaco]

Men, if considering psu can pop flames after few years of use because they have caps, then any equipment should be of suspect that contains caps

 

[wingman99]

Sweet jebus.......

So are you telling me, wingman, that you would bet on a new PSU failing before a 13 year old would? That is the only point I am trying to make here... nobody said anything about it catching fire... just a siple failure. But surely it's more likely a 13 year old PSU will fail before a brand new one.

RE: The EVGA GPUs... I recall one, maybe two/three of those EVGA GPUs catching fire.. don't believe everything you read... Its not like a dozen or a hundred even did so.

This is about preventing PSU fires, do you need to replace a old PSU prevent fires. My point was only few EVGA GPUs caught fire and the internet caught the few that did. If it is dramatic device catching on fire it is on the net. So where is all the 10+ year old PSUs catching fire on the internet. I searched in google for psu catching fire found these two.

My pc just randomly turned off, completely shut down. I smelt a smell of burning so I unplugged the pc straight away.

I turned it back on and smoke started coming out from my PSU. I turned it off immediately.



I turned on only the PSU next, so I could fully see what was happening. With only the motherboard power connectors in place so I could turn it on.



I saw inside a flame inside, so I turned it off immediately.



It's EVGA 750 G2 Gold power supply

It's not old either, I only recently built this pc, only around 6 months ago. It's gfx card is a GTX 970 FTW+ Edition. https://linustechtips.com/main/topic/481360-psu-caught-fire/

EVGA PSU lit my house on fire.

Does anyone have any idea why my 1 year old EVGA 450W power supply went up in flames while my computer was in sleep mode?

If it helps the specs of the computer was:

-asus rog strip b450-f mother board

-1050 TI graphics card

-gen 2 ryzen 3 cpu

-16g ripjaw 3200 ddr4

- Corsair 460x RGB

-all stock coolers, nothing fancy, nothing overclockedhttps://linustechtips.com/main/topic/982932-evga-psu-lit-my-house-on-fire-🔥/

 

[Alaric]

I'll try less doom and gloom and go with pragmatism. Top racing classes in the NHRA, NASCAR, Formula 1, etc., change major engine and chassis parts after a predetermined number of laps/miles/races/hours/whatever. They determined that a crankshaft, for example, has a probability of failing after 10 races. So they are replaced after 7 races with the Murphy's Law fudge factor figured in. It's still a good part, yet it's thrown in the trash because the cost of that part failing is much greater than the cost of replacing wear items. That's how they get to be wear items.

Which is why I bought a new power supply when the old on still worked. Wagering 20 times the value of my PSU would have been dumb. The old PSU is GOING to fail, it's a matter of when. The new one doesn't have those extra years on it so logic dictates that my odds are better with the new one, and the new one was 1/20 the cost of the old one failing. The people here mostly build custom, expensive rigs. If you want to run off the shelf rigs that are built to a price point, buy a Dell. They're made to throw out when something breaks. High performance equipment costs more to maintain, there is no such thing as a free lunch. Speed costs money, how fast do you want to go?

To clarify, Dells are built like freakin' tanks. I wasn't disparaging the brand. It was an example of "stock" vs. the rigs common to OCF.

 

[Niku-Sama]

by that thinking how many "races" does a PSU have in it untill murphy fudge warrants a change out?

 

[Alaric]

Since it costs considerably less than a $5000 crankshaft, I never checked. LOL I've used old parts, and I'm sure I will again some time. When someone asks for advice, though, I'll give what I believe to be the best practice answer. Individuals are responsible for their own budgetary issues. We start at the top with "best answer" and work our way down the list until one has "I can afford it" next to it.

 

[EarthDog]

This is about preventing PSU fires, do you need to replace a old PSU prevent fires. My point was only few EVGA GPUs caught fire and the internet caught the few that did. If it is dramatic device catching on fire it is on the net. So where is all the 10+ year old PSUs catching fire on the internet. I searched in google for psu catching fire found these two.



EVGA PSU lit my house on fire.

???

This isnt about psu fires. You are coming across like when these fail, they all catch fire. Things tend to fail without fireworks most of time.

After 13 years, I'd bet on it failing before anything new would. Not spectacularly with flames... just...stop... working. This isnt about fires (something incredibly rare when talking psu failures)...just failures.

 

[UltraTaco]

Well, I guess new isn't so bad, look what I found!

https://www.ebay.com/itm/680-WATT-6...483378&hash=item41728af93c:g:yOoAAOSwZqZaL5gz

 

[EarthDog]

Yes taco, it needs to be a QUALITY unit (unlike what you linked likely). I assumed that was a given here but, apparently it needs to be spelled out...

 

[Robert17]

Here is some insight from a tech paper and a PC vendor. The tech paper has some math which I'm sure Taco will max out with his slide rule. The PC vendor's info is from 2016 but is the most recent I could find. In conclusion, don't let them get hot.

A simile to Alaric's post: I change my car's oil before the engine seizes up. Regular consideration of replacement parts like oil filters go along with it. Cheaper than a new engine.

 

[EarthDog]

Robert...awesome link! I used to refer to hardware.fr who used to do this but stopped a couple years back... this is great!

So... 3.11 percent fail in the first year. Anyone taking bets on if those same psus are 13 years old that the failure rate increases due to wear failures at that age? I mean it's not a leap of faith (for most) to think it does...

Edit: Robert's first link shows exactly that. It's wear failure that go up exponentially as time goes on. So yeah, there are always doa's and 'infant' mortality, but as time goes on, it's clear failure rates increase.

After 13 years...past its operational life....

The final phase is the “wear-out” phase, where
the supply fails as its components reach the end
of their operating life. Common mechanisms for
wear-out include fan bearings going bad, electrolytic
capacitors drying out, and stress cracks developing
after thousands of thermal cycles.

 

[freeagent]

Yup, and if there are ever any doubts, pop it open for a quick inspection and cleaning, if anything I'm sure there are some full on rabbits tucked in some crevices if its been a few years since the last one.. My oldest psu is an old TX850 that I beat pretty hard with x58 and a pair of 570s benching and folding. She still works.. but in standby on a shelf. I gave my buddy an old Neo 480 from the A64 days and its running an A64 4400. Classic. It might even be my old 4400, I have no idea what I did with it.

 

[EarthDog]

Please do not pop open a psu and clean it... it can kill you. Not a good idea if you dont know what you are doing.

 

[Alaric]

Please do not pop open a psu and clean it... it can kill you. Not a good idea if you dont know what you are doing.

If you just have to, though... Unplug it from the wall and press the power button on the computer. That should-should-drain the caps. Let it sit for 48 hours after that. This in no way guarantees that you can't get zapped, but it certainly swings the odds to more in your favor. Personally I would prefer to plug it in, short the green wire to a black wire on the ATX plug, and just blow some compressed air through the inlet.

 

[freeagent]

Yeah I just meant pop it open to inspect the caps.. ie look for bulging, leaking, burnt marks etc. As for cleaning, compressed air should be fine no need to be hardcore and break out the q tips.. unless you want to give the fan a scrape

 

[UltraTaco]

Taco opened one up a while back, still alive be gentle and no touching.

 

[Zerileous]

Back in 03 it was said (after unplugging and using the power button a few times), to bridge the blades on the outlet plug with a screwdriver, and then let it sit for 48 hrs. Well at the time it sounded safe and I'm still here. I think I just *had* to take a Dremel to every restrictive punched steel fan grille I could find, including the PSU. Of course removing the contents from the enclosure first was advisable.

I really liked the bathtub graph in the first link Robert posted, seems to explain the anecdotal evidence we see here. I do question the choice of "Infant mortality" to describe early failures? Is that a common term in the industry or just editorial expression? We all have different experiences that shape our perception of phrases like this. While it doesn't particularly bother me, it does strike me as quite morbid. However, I could see how an engineer might find it to sound clever.

I see nothing wrong with the eBay power supply. The used black paint after all, so it must be good!

Oh yeah, and just replace it!

 

[Alaric]

I change my car's oil before the engine seizes up. Regular consideration of replacement parts like oil filters go along with it. Cheaper than a new engine.

I've seen the inside of a LOT of blown up engines. Never saw one fail because the oil was too clean. LOL

 

[UltraTaco]

Alaric, sir, oil doesnt need to be clean! As long as theres oil in there, all is fine. You drive 10 minutes and oil is black already after it's been changed..