Running a 1000W PSU for 13 years now. Should I replace it or wait for it to fail?

[jaslion]

On 11/17/2021 at 6:19 AM, dilpickle said:

Who said anything about 15+?

 

But if you want to pay for some plane tickets I'll show you even older hardware.

And I have a gba who's ac adapter (aka power supply) stopped working last year. But also a pentium 3 win xp old game box that is going on 20 year old hardware and that was used as the main pc in the house for almost 14 years.

 

So yes stuff just randomly goes after a while and you cannot guarantee it lasts. Hence why you see those 2,3,5,7 or 10 year psu warranties because that is the most they can safely estimate for how long the psu will work under normal predicted use.

 

You can not say that stuff will keep working that is just flat out a wrong train of thought. Everything breaks and the older something gets the higher chance of it. This is why businesses usually flat out replace after a 3-5 year period.

[Alder Lake]

On 11/16/2021 at 5:12 AM, tulx said:

Hello everyone,

 

I´ve been running a Zalman ZM1000-HP as the only constant component in my various builds since 2008 and, apart from some slightly annoying coil whine, it´s been working perfectly. I´m beginning to wonder now if I should proactively replace it at some point. I think the PSU is rated at 100.000 hours and I could be at about 75.000 hours accoring to a very generous estimation. I´d have no issue with running it until it fails in another 10 or so years if had the certainty that it would not take my other components with it when it eventually does. Are there any experiences with similarly old components or a general best practice for such cases?

Thanks a lot in advance for your time and have a pleasant week!

 

Kind regards,

tulx 

might want to buy another one purely for the coil wine

 

you could've probably asked about warranty coverage over coil wine, they might have given you another one

[dilpickle]

4 hours ago, jaslion said:

 

You can not say that stuff will keep working

When did I say that

[Mister Woof]

From my experience, businesses and government agencies tend to not worry about actual failure and focus on warranty/contract length regarding upgrades.

 

They view it all as an ongoing business expenditure, and budget for them based on those two metrics. This is why we so often see boatloads of old SFF systems for cheap from wholesalers because they go out of warranty/contract, and have another contract to replace them all. 

 

Will these PSUs last longer than their warranty length? Of course, but as far as the business is concerned, they never reach that date.

 

The reason you see them in schools a lot is because they are often underfunded and simply have zero budgets for computers. They get donations from businesses that can help with tax write-offs and also hand-me-downs from government agencies.

 

Then again, these systems are far, FAR from high-end components. They won't see the load types that warrant exceptional protections. Not like what you'd want from a 1000w PSU. 

 

And more importantly, since they're purchased in bulk...so what if they fail? Nothing local is generally important, and if it fails, they simply pull another unit from the pallet. And have service contracts from their vendor during the length of their warranty period.

 

You as a user generally don't have that type of support. If your 10 year old 1000w PSU dies and takes the whole system with it, you don't have a pallet of clones to simply replace the entire unit.

[Bad5ector]

On 11/16/2021 at 9:50 PM, dilpickle said:

Jesus is this thread some viral marketing for hardware manufacturers? Who the hell "proactively" replaces parts while they are still working?

I for one would. If I suspect a faulty PSU in my main rig, I don't hesitate to replace. I'll just continue to use the old part in another less important build and run it to the ground.

[dilpickle]

1 hour ago, Mister Woof said:

 If your 10 year old 1000w PSU dies and takes the whole system with it

Honestly what do you think the odds of that happening are? How many cases have you seen? It's just as likely that a brand new unit could also take out your whole system. I don't see what you can do about that. Other than make backups.

 

The OP has a brand name unit with all the same safety features as one made today. Its been running for 13 years so obviously it is not defective. If it was it would have failed a long time ago. The worst case scenario is one day it will die and it gets replaced.

 

I am not aware of some revolution in power supply design in the last 10 years so maybe you can enlighten me here.

[Mister Woof]

24 minutes ago, dilpickle said:

Honestly what do you think the odds of that happening are? How many cases have you seen? It's just as likely that a brand new unit could also take out your whole system. I don't see what you can do about that. Other than make backups.

 

The OP has a brand name unit with all the same safety features as one made today. Its been running for 13 years so obviously it is not defective. If it was it would have failed a long time ago. The worst case scenario is one day it will die and it gets replaced.

 

I am not aware of some revolution in power supply design in the last 10 years so maybe you can enlighten me here.

Why even have protections at all?

 

 

[Dedayog]

26 minutes ago, dilpickle said:

Honestly what do you think the odds of that happening are? How many cases have you seen? It's just as likely that a brand new unit could also take out your whole system. I don't see what you can do about that. Other than make backups.

 

The OP has a brand name unit with all the same safety features as one made today. Its been running for 13 years so obviously it is not defective. If it was it would have failed a long time ago. The worst case scenario is one day it will die and it gets replaced.

 

I am not aware of some revolution in power supply design in the last 10 years so maybe you can enlighten me here.

Incorrect, the worst case is that it can fail and catch fire and he loses his PC, house and family in a bonfire that was his home.  I mean, if we want to go worst cases.

 

But we are arguing the fact that it may fail and take other components with it. That's what prompted a lot of this, and my questions previously.

 

If it just failed and then done, this would be a non-issue.  

[dilpickle]

9 minutes ago, Dedayog said:

Incorrect, the worst case is that it can fail and catch fire and he loses his PC, house and family in a bonfire that was his home.  I mean, if we want to go worst cases.

 

 

Again that can happen with a brand new unit. In fact it can happen with any electronic device you've ever bought. So what do you suggest we do?

[Dedayog]

6 minutes ago, dilpickle said:

Again that can happen with a brand new unit. In fact it can happen with any electronic device you've ever bought. So what do you suggest we do?

I suggested a few things in my prior post, asking some questions as well.

 

But you're missing the other side.  Warranty.  During warranty, that's your safe coverage feel good.  Once a unit is out of warranty, then we get into this discussion.

 

Do people feel safe using a PSU after it has aged out of warranty, with nothing to safeguard them at all.

 

I don't have an answer but still think that's the crux of it.  Should units have an end of life setting when warranty is over, for example?  Or just do a hope and fingers crossed that when it finally dies it's either out of your system, or just dies quiet?

 

As things stand now we just do fingers crossed and that is that?  Since we can't know how it will end, we take our chances.  I am asking if there is another alternative.

[Mister Woof]

8 minutes ago, Dedayog said:

I suggested a few things in my prior post, asking some questions as well.

 

But you're missing the other side.  Warranty.  During warranty, that's your safe coverage feel good.  Once a unit is out of warranty, then we get into this discussion.

 

Do people feel safe using a PSU after it has aged out of warranty, with nothing to safeguard them at all.

 

I don't have an answer but still think that's the crux of it.  Should units have an end of life setting when warranty is over, for example?  Or just do a hope and fingers crossed that when it finally dies it's either out of your system, or just dies quiet?

 

As things stand now we just do fingers crossed and that is that?  Since we can't know how it will end, we take our chances.  I am asking if there is another alternative.

That just opens the door for artificially planned obsolescence which is the kinda shit Apple does and sort of ties in with right to repair.

 

So from that perspective, I hope that never happens.

 

Everyone just needs to make their own good judgments and risk assessment.

[Guest]

2 hours ago, dilpickle said:

Honestly what do you think the odds of that happening are? How many cases have you seen? It's just as likely that a brand new unit could also take out your whole system. I don't see what you can do about that. Other than make backups.

Given I'm in the industry, I see it all of the time.  At least once a week, a box comes in with a damage claim and we have to do failure analysis.  But that's once a week when tens of thousands of units are sold.  So odds are pretty slim, but the odds are still there.  Because the odds are not good to win the lottery, we do we then choose not to play?

 

And I'm going to have to break everyone's heart here and break it to you all that desktop components, when compared to server grade, military grade or medical grade, are complete SHIT.  Yes.  Even your precious Seasonic Prime and Corsair HX1200.   Power supplies used to be very well made most of the time (don't get me wrong, they were also a lot worse too, but fortunately safety regulations and lawsuits weeded most of those out), but then people started to ignore the role of a power supply and started saying "just give me a 750W".  When you develop a PSU product, you can't just say "I'm going to make this the best whatever" and go from there.  You start by looking at what else is sold on the market, and at what price and then try to make it as good as possible while staying within a price range that people are willing to pay.

 

We can make a power supply that's virtually indestructible, that is 100% safe, that outperforms everything else on the market.... but it will cost 3x as much and nobody will buy it.

 

I live in the town next to the one that has the world's longest burning light bulb.  They call it the Centennial Light.  It's been burning in the Livermore, CA fire station since 1901.  The bulbs were made by Shelby Electric.  The company didn't last long before going bankrupt and getting bought by GE.  Why did the manufacturer of the best light bulbs go out of business?  Because the bulbs were too expensive compared to GE's even though they lasted 100 times longer and they didn't get repeat business because their existing customers didn't need to replace their Shelby bulbs.  Your PC power supply is a GE. Not a Shelby.

[Dedayog]

14 minutes ago, jonnyGURU said:

Given I'm in the industry, I see it all of the time.  At least once a week, a box comes in with a damage claim and we have to do failure analysis.  But that's once a week when tens of thousands of units are sold.  So odds are pretty slim, but the odds are still there.  Because the odds are not good to win the lottery, we do we then choose not to play?

 

And I'm going to have to break everyone's heart here and break it to you all that desktop components, when compared to server grade, military grade or medical grade, are complete SHIT.  Yes.  Even your precious Seasonic Prime and Corsair HX1200.   Power supplies used to be very well made most of the time, but then people started to ignore the role of a power supply and started saying "just give me a 750W".  When you develop a PSU product, you can't just say "I'm going to make this the best whatever" and go from there.  You start by looking at what else is sold on the market, and at what price and then try to make it as good as possible while staying within a price range that people are willing to pay.

 

We can make a power supply that's virtually indestructible, that is 100% safe, that outperforms everything else on the market.... but it will cost 3x as much and nobody will buy it.

 

I live in the town next to the one that has the world's longest burning light bulb.  They call it the Centennial Light.  It's been burning in the Livermore, CA fire station since 1901.  The bulbs were made by Shelby Electric.  The company didn't last long before going bankrupt and getting bought by GE.  Why did the manufacturer of the best light bulbs go out of business?  Because the bulbs were too expensive compared to GE's even though they lasted 100 times longer and they didn't get repeat business because their existing customers didn't need to replace their Shelby bulbs.  Your PC power supply is a GE. Not a Shelby.

As you say, it also applies to every other thing we buy as well.  Good reminder.

[Mister Woof]

13 minutes ago, Dedayog said:

As you say, it also applies to every other thing we buy as well.  Good reminder.

Yeah, but to a certain extent it would just be unnecessary cost.

General user doesn't need some super heavy duty transmission for their grocery getter car - the ones we use are rated to be used at x/y/z power and miles and they are more than enough for the balance of price/performance. Better to divert those manufacturing costs towards things the user will care about, like creature comforts and features.

Same goes for stuff like rifle barrels - people go on and on about making sure to have chrome lined CHF barrels as some kind of end-all....nothing against C/L CHF barrels, but your guy perusing PSA cheap-o nitride barrels ain't gonna shoot the thousands and thousands of rounds needed to make a much more expensive barrel be worthwhile. Better to get the cheaper nitride 4140 and buy more ammo.

 

In the end every part has a rated limitation, and as long as we are within those limitations and have the right expectations, we're doing okay.

[dilpickle]

57 minutes ago, jonnyGURU said:

Given I'm in the industry, I see it all of the time.  At least once a week, a box comes in with a damage claim and we have to do failure analysis.

And what were the results of those analysis? So you're saying once a week you get a computer with a blown power supply that destroyed every component in the system? I find that very hard to believe.

[Sleepycat3]

1 hour ago, jonnyGURU said:

And I'm going to have to break everyone's heart here and break it to you all that desktop components, when compared to server grade, military grade or medical grade, are complete SHIT.  Yes.  Even your precious Seasonic Prime and Corsair HX1200.   Power supplies used to be very well made most of the time (don't get me wrong, they were also a lot worse too, but fortunately safety regulations and lawsuits weeded most of those out), but then people started to ignore the role of a power supply and started saying "just give me a 750W".  When you develop a PSU product, you can't just say "I'm going to make this the best whatever" and go from there.  You start by looking at what else is sold on the market, and at what price and then try to make it as good as possible while staying within a price range that people are willing to pay.

I'd really love your insight on this as I have been reading through your evaluations of PSUs on your old site over the past 5-8 years. Was there a period when PSUs switched over from being "over designed" for durability and reliability to "lowest ripple, highest efficiency but last just the warranty duration"?

 

I'm now running my 2007/2008 HX850W in my 5900X / 3080 and it hasn't skipped a beat surprisingly. So I'm wondering based on your experience and what you have seen, whether there was a step change where profits were heavily prioritised over product engineering.

Edited November 18, 2021 by Sleepycat3

[Guest]

1 hour ago, dilpickle said:

And what were the results of those analysis? So you're saying once a week you get a computer with a blown power supply that destroyed every component in the system? I find that very hard to believe.

Well... not EVERY component.  But maybe the motherboard and/or graphics card.  Sometimes just a hard drive.  And that's once a week it's an actual PSU fault.  Catastrophic failures on the secondary side are rare, but when they do happen, they're typically spectacular.  Like when an SR MOSFET suffers thermal runaway and blows up.  Protections don't typically catch that because it happens so fast.

 

Over all, we probably get a dozen a week that are "claims",  but then we find out that the fault is caused by user error, like wrong cables used or too long of a mounting screw tapping into the PCB, etc.

 

But, like I said, we're talking about one out of probably around 50K, so it's a very, very small percentage.  Most of the time when a PSU dies, it just dies.

 

1 hour ago, Sleepycat3 said:

I'd really love your insight on this as I have been reading through your evaluations of PSUs on your old site over the past 5-8 years. Was there a period when PSUs switched over from being "over designed" for durability and reliability to "lowest ripple, highest efficiency but last just the warranty duration"?

 

I'm now running my 2007/2008 HX850W in my 5900X / 3080 and it hasn't skipped a beat surprisingly. So I'm wondering based on your experience and what you have seen, whether there was a step change where profits were heavily prioritised over product engineering.

Strives for efficiency, new safety requirements, desire for smaller size, and lower cost isn't a new thing.  When I say "they used to be better", think about the time when we had PC Power & Cooling Turbo Cool, Zippy EMACS, Impervio and even some Delta made units like Antec HCP Platinum.

[Dedayog]

2 hours ago, dilpickle said:

And what were the results of those analysis? So you're saying once a week you get a computer with a blown power supply that destroyed every component in the system? I find that very hard to believe.

Not sure if you're aware...

 

This is the guy posting.  May lend some weight to his arguments.

 

johnny guru - Search (bing.com)

 

[Guest]

1 hour ago, Dedayog said:

Not sure if you're aware...

 

This is the guy posting.  May lend some weight to his arguments.

 

johnny guru - Search (bing.com)

 

Bro... Spell my name right. 

 

[Dedayog]

16 minutes ago, jonnyGURU said:

Bro... Spell my name right. 

 

You are legitimately too old to use "bro"    More like "Fella"

 

[Guest]

At least you didn't say "Ok boomer."

 

[Cifer]

9 hours ago, jonnyGURU said:

Strives for efficiency, new safety requirements, desire for smaller size, and lower cost isn't a new thing.  When I say "they used to be better", think about the time when we had PC Power & Cooling Turbo Cool, Zippy EMACS, Impervio and even some Delta made units like Antec HCP Platinum.

Cooler Master's V Platinum lineup is supposed to be the same thing, but I never see people recommending it.
Do you know of something problematic with Cooler Master's implementation?

[Guest]

9 hours ago, Cifer said:

Cooler Master's V Platinum lineup is supposed to be the same thing, but I never see people recommending it.
Do you know of something problematic with Cooler Master's implementation?

Might be fine.  I know nothing about it.

 

[tulx]

Hello everyone and thanks a lot for the numerous insightful replies! Since my other components relying on the 13 year old PSU are quite a lot newer (5900X, a 3080 and a water cooling loop), I will follow what seems to be the general consensus and buy a new PSU - would be a shame to lose the newer much more expensive components, even if the chance of that happening is low.

I consulted the PSU tier list here and seems that something like a be quiet STRAIGHT POWER 11 1000W would be a good fit for the next decade. 

[tulx]

Installed the new be quiet PSU and am amazed that it has no coil whine at all. After 13 years of basically being able to play tunes by moving around Explorer windows or the camera in games, this is pretty groundbraking! Thanks again for the advice, folks!