7 year old PSU?

[sandro27]

I have had an InWin p85 power supply since 2019 and am wondering if there is any safety concern with it given its age and that it has been used pretty consistently for its whole lifespan. I can guess it has around 5000 power on hours based on a drive of the same age in the same system. Anyone have any info?

[Hinjima]

6 minutes ago, sandro27 said:

I have had an InWin p85 power supply since 2019 and am wondering if there is any safety concern with it given its age and that it has been used pretty consistently for its whole lifespan. I can guess it has around 5000 power on hours based on a drive of the same age in the same system. Anyone have any info?

Its an okay power supply and if it has lasted this long is should last a bit longer.

It does only come with a 5 year warranty so that is out the window, higher quality PSU's come with 10 year warranty.

 

Its a tough one,  I would replace it if you are considering new hardware as well.

It can fail tomorrow and it can fail in 8 years, non of us are able to tell you this.  I would feel safer with a PSU with warranty.

[leclod]

A decent psu shouldn't take other parts with it when it fails.

My psu is 10 1/2 years old and I don't plan to replace it.

Edited March 14 by leclod

[porina]

I usually use PSUs until they fail, which outside of two specific cases, is practically never. They get obsolete or sold with old systems before they fail. The oldest PSU I still have in active use is from 2017, and two more from 2019. I have no intention to replace any of them because of age.

 

The two cases of failed PSUs are:

1, An MSI unit that failed within warranty. It got replaced by same new and that also failed within warranty. Seller replaced with totally different brand so hopefully that wont happen again.

2, Maybe 20 years ago, there was a "400W" Gold PSU - yes, they were metallic yellow coloured. 80Plus didn't exist back then. Everyone knew they weren't 400W but it was so cheap no one cared. They lasted about a year 24/7 running before dying. Then you buy another one. Never killed any attached hardware. If you wonder why not buy a more expensive one that didn't die, this was back when electricity was cheap enough that many people ran barebones crunchers, and keeping the cost per node minimal cost made a difference to scaling.

[Kisai]

19 minutes ago, leclod said:

A decent psu shouldn't take other parts with it when it fails.

My psu is around 10 years old and I don't plan to replace it.

Oh, they will.

 

It depends on the load when it "dies". If it dies on start up, it will usually not kill anything else. If it "dies" when the GPU spins up, it can cause the thermal solution to be unable to cool the GPU/CPU/SSD/etc if the cooler's mass is insufficient. 

 

Basically, standard Overcurrent protection preventing it from turning on, or turning it off when overcurrent is detected is correct operation. Incorrect operation is when connectors thermally expand due to wires being the wrong gauge for the current *cough*12VHPWR*. 

 

I have NEVER seen a (continually running) working PSU fail. I've only seen new/new-old-stock PSU's fail, or PSU's in homes where the customer smokes ever fail, and they've generally just had the fans fail.

 

[leclod]

4 minutes ago, Kisai said:

Oh, they will.

You mean a decent psu can take other parts with it when it fails ?

it can, yes but will it ? if it fails at all ?

to me chances are pretty low, low enough to risk it.

[Kisai]

Just now, leclod said:

You mean a decent psu can take other parts with it when it fails ?

it can, yes but will it ? if it fails at all ?

to me chances are pretty low, low enough to risk it.

The risk is low enough if you bought a good enough PSU (eg Seasonic) and you're only using the PSU's supplied cables. If you're using third party cables, or liquid cooling, or undersized coolers, then you increase the risk a few points as all of these increase the risk.

 

Liquid coolers are more likely to leak or fail than the PSU.

[Needfuldoer]

[ Moved to Power Supplies ]

 

If it was a good power supply when it was new, and you know its history (it wasn't run hard in extreme conditions), it's probably fine. 

 

10 years isn't that long for parts like power supplies. If it was a Seasonic or higher-end EVGA, it wouldn't even be out of warranty yet!

[Beerzerker]

I'm running a 10 year old unit in this one right now - No problems to mention at all. 

[PDifolco]

3 hours ago, Kisai said:

Oh, they will.

 

It depends on the load when it "dies". If it dies on start up, it will usually not kill anything else. If it "dies" when the GPU spins up, it can cause the thermal solution to be unable to cool the GPU/CPU/SSD/etc if the cooler's mass is insufficient. 

 

Basically, standard Overcurrent protection preventing it from turning on, or turning it off when overcurrent is detected is correct operation. Incorrect operation is when connectors thermally expand due to wires being the wrong gauge for the current *cough*12VHPWR*. 

 

I have NEVER seen a (continually running) working PSU fail. I've only seen new/new-old-stock PSU's fail, or PSU's in homes where the customer smokes ever fail, and they've generally just had the fans fail.

 

Did you see any vid/story of a PSU killing things while dying ?? Looks sus to me, when the PSU dies everything stops and CPU/GPU temps drop drastically (as heat is on small areas due to electricity passing through) 

What could be dangerous is PSU dying "dirty" with flames that can burn a whole house - but again it's way rarer than with a 12VHCrap plug or a phone/laptop battery !  (for which I have irl experience, my neighbor's flat was ravaged by a laptop induced fire and we were lucky it didn't spread more) 

[Toryze]

19 hours ago, Kisai said:

Oh, they will.

 

It depends on the load when it "dies". If it dies on start up, it will usually not kill anything else. If it "dies" when the GPU spins up, it can cause the thermal solution to be unable to cool the GPU/CPU/SSD/etc if the cooler's mass is insufficient. 

 

Basically, standard Overcurrent protection preventing it from turning on, or turning it off when overcurrent is detected is correct operation. Incorrect operation is when connectors thermally expand due to wires being the wrong gauge for the current *cough*12VHPWR*. 

 

I have NEVER seen a (continually running) working PSU fail. I've only seen new/new-old-stock PSU's fail, or PSU's in homes where the customer smokes ever fail, and they've generally just had the fans fail.

 

Whether or not a PSU kills other components in the event of a catastrophic failure entirely depends on what fails, where within the circuit, and how it fails. If the failure occurs on the Primary side (the circuit before it enters the transformer), the risk to connected components is essentially zero, due to galvanic isolation from the transformer. Optocouplers and/or digital isolator ICs also electrically "separate" the primary and secondary side of the PSU but allow for communication.

If it occurs on the secondary side, it's a lot more... difficult. You may be lucky; you may be not.

For 12VHPWR/-2x6, the wire gauge is rarely the problem itself. It's primarily the terminals used, alongside the finnicky design and terrible specification for the expected current, not to mention the possibility of poor QC that doesn't catch bad crimps if they occur. It's a rather poor design for handling its load within the specification and has numerous shortcomings, unfortunately.

It's not unexpected to see PSUs working fine for most PCs/loads, even the more poorly designed units, mainly due to the lack of stress on the power supply. Since the fan is the only moving mechanical part, it's basically understood it'll be the first to go in an ideal manner. If there's a death before a natural longer-term fan failure, it's generally due to some factor of improper QC or shipping damage making its way to the consumer.

20 hours ago, Kisai said:

The risk is low enough if you bought a good enough PSU (eg Seasonic) and you're only using the PSU's supplied cables. If you're using third party cables, or liquid cooling, or undersized coolers, then you increase the risk a few points as all of these increase the risk.

 

Liquid coolers are more likely to leak or fail than the PSU.

Whether or not the circuit of the PSU was properly designed matters a lot more than exactly whoever designed it.

Moving parts will eventually fail, yes. Especially if they're subjected to higher stresses.