Is using 10-years old power supplies a concern?

Napoleon

[H]ard|Gawd
Joined
Jan 27, 2003
Messages
1,071
not a concern, but IMO, the nicest thing you can do for the rest of your components is to make sure the supply is putting out consistent voltages. My theory isn't The better power inputs they get the less their own voltage hardware has to work.

I replaced a corsair tx750 from 2009 with an EVGA p2750 and wow, voltage drop and fluctuation is almost nothing using the same hardwaremonitor software.
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
11,358
Age itself isn't really the biggest variable. What kind of life has the power supply had? Age becomes a concern mainly when a power supply has had a rough life. Did it run hot inside a case that had poor airflow and/or never got dusted out? Was the power supply pushed to it's limits on a regular basis? It wouldn't be unreasonable to have a PSU last 30+ years if it is treated well. A PSU that has been abused or neglected would be lucky to last a handful of years.
 

Dead Parrot

2[H]4U
Joined
Mar 4, 2013
Messages
2,831
Another thing to consider is how messed up is the input power? Frequent drops, spikes and noise can take a toll on the input filters.

If you can, blow all the fuzz and crap out of them and do a visual inspection. Look for bulging or leaking caps, burned areas etc. If they pass the visual and have had a fairly easy life, they could well last another 10 years.
 
Joined
Mar 29, 2017
Messages
61
If there are no spikes and fluctuations even a 20 years old PSU will work fine, provided it supports the system.
PS: I am using 13 years old PSU with my Core2duo.
 

GiGaBiTe

2[H]4U
Joined
Apr 26, 2013
Messages
2,398
Ten years is pretty long in the tooth for a power supply. While all of the solid state components will likely be fine, the capacitors are going to start showing their age. I would expect the capacitors to have higher than normal ESR and may start to become electrically leaky, which stresses other components and can cause regulation to go out of spec. If you want continued reliable service out of your computers, I would recommend changing them out.

It wouldn't be unreasonable to have a PSU last 30+ years if it is treated well. A PSU that has been abused or neglected would be lucky to last a handful of years.

A sealed PSU is never going to last 30 years of service, no matter the usage, without being completely recapped at least a couple of times. Head on over to the vintage computer forum where we have PSUs from the mid 90s and older with dead electrolytics and exploding tantalum caps.
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
11,358
A sealed PSU is never going to last 30 years of service, no matter the usage, without being completely recapped at least a couple of times. Head on over to the vintage computer forum where we have PSUs from the mid 90s and older with dead electrolytics and exploding tantalum caps.

So a few examples of PSUs that have probably been treated like shit means others could never last longer? I have several examples of 286, 386, and 486 computers that all still work fine and have never been recapped, though granted they don't see much usage these days, they do still work. I have a full audio rack full of gear that is over 40 years old, what makes you think a computer PSU couldn't last 30? Some of us actually have good airflow through our computers and don't just let our PSUs cook. Both of the 1000w PSUs in my 5820k and 2500k rig date back to ~2007 and both still look brand new inside and work perfect.
 

sirmonkey1985

[H]ard|DCer of the Month - July 2010
Joined
Sep 13, 2008
Messages
22,231
yeah i have a thermaltake TP 750w that's over 10 years old as well.. other than some dust everything in it looks perfect.. if you treat it right and it's a well made psu they'll last a long time.
 

Nenu

[H]ardened
Joined
Apr 28, 2007
Messages
20,223
I'm using a Corsair AX750 that has finally exceeded its 7 year warranty by 1/2 a year.
Its still like new and has been running 24/7.
Despite the initial outlay it has easily paid for itself over that period and looks like its got a lot left.
The peace of mind I still have alone is worth the money.

Mine is also a Seasonic under the hood.
Quality PSUs are worth it.

edit
ah scratch the time info, its 6.5 yrs old.
 

GiGaBiTe

2[H]4U
Joined
Apr 26, 2013
Messages
2,398
So a few examples of PSUs that have probably been treated like shit means others could never last longer? I have several examples of 286, 386, and 486 computers that all still work fine and have never been recapped, though granted they don't see much usage these days, they do still work.

Yes, and a car with cylinder misfires can still run down the road. It matters not if a capacitor was run smokin' hot, babied or never run at all, they have lifetime ratings because the electrolyte and internal structure of the capacitor breaks down even without use. Why do you think old wax and paper capacitors from the early days of radio and TV sets are nuking vacuum tubes? The paper goes acidic and effectively turns the capacitor into a resistor and passes DC across them. The same happens with more modern electrolytics, it just takes longer.

I have a full audio rack full of gear that is over 40 years old, what makes you think a computer PSU couldn't last 30? Some of us actually have good airflow through our computers and don't just let our PSUs cook. Both of the 1000w PSUs in my 5820k and 2500k rig date back to ~2007 and both still look brand new inside and work perfect.

"My audio gear looks fine and makes noises so it's OK" is not a valid testing methodology of capacitors. Either get out an LCR meter and prove all of your 40 year old caps don't have internal electrical leakage, out of spec ESR or other problems, or stop trying to say capacitors never have age related failure. You can polish a turd all you want, it's still going to be a turd no matter how glossy it gets.
 

Gavian

Limp Gawd
Joined
Jan 13, 2016
Messages
437
Wouldn't worry about it if it's a reliable brand.

I still have a 12 year old Silverstone Zeus 650W PSU in my secondary PC, it outlived every other component from the original build in 2005.
 
Last edited:

Araxie

Supreme [H]ardness
Joined
Feb 11, 2013
Messages
6,456
I'm still running great a 11 years old Thermaltake ToughPower W0172 850W non-modular, which has been tortured for years.. I just do the periodic cleaning every 5 - 6 months, capacitor check every year and everything still flawless with that PSU. it's nice to remember that PSU was made before the Eco's "80 Plus" certification.
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
11,358
Yes, and a car with cylinder misfires can still run down the road. It matters not if a capacitor was run smokin' hot, babied or never run at all, they have lifetime ratings because the electrolyte and internal structure of the capacitor breaks down even without use. Why do you think old wax and paper capacitors from the early days of radio and TV sets are nuking vacuum tubes? The paper goes acidic and effectively turns the capacitor into a resistor and passes DC across them. The same happens with more modern electrolytics, it just takes longer.

"My audio gear looks fine and makes noises so it's OK" is not a valid testing methodology of capacitors. Either get out an LCR meter and prove all of your 40 year old caps don't have internal electrical leakage, out of spec ESR or other problems, or stop trying to say capacitors never have age related failure. You can polish a turd all you want, it's still going to be a turd no matter how glossy it gets.

I am a longtime member of the website Audiokarma.org which largely focuses on vintage audio 30-60 years old. All of my 40 year old gear HAS indeed been gone over by qualified technicians, and aside from some cracked solder joints, not found to have any issues with their capacitors. You have to go way back past 40 years to get to where wax and paper capacitors were still common, none of my units from the 70's have those.

I at no point said that there can't be age related failures. What I said is "Age itself isn't really the biggest variable", which is true. Certainly there can be failures based on age, but that is where quality and care comes into play. Plenty of cheap power-supplies are made with cheap capacitors and if that is what you run, you deserve what you get. One amazing thing about my gear with quality capacitors from the 70's still working 40 years later is that the quality capacitors in 10 year old electronics are 30 years newer, with 30 years of improvements - to say that they should work even better is an understatement. To be concerned about age with these after only 10 years is crazy.

And how the fuck did you actually post that "It matters not if a capacitor was run smokin' hot, babied or never run at all"? Did you actually type that with a straight face? Temperature is absolutely the #1 thing that helps determine the life of most electronics, especially if we are talking decades.
 
Last edited:

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
54,075
I am still using a couple of power supplies that my father bought 10 years ago.

Is that a concern?

Both of the units still work flawlessly and I haven't encounter any issues.

The said units are Corsair TX650W built by Seasonic

http://www.silentpcreview.com/article813-page1.html

https://outervision.com/articles/psu/corsair/tx650/corsair-tx-650-1
This might be of interest for you.

SilverStone Olympia 1000W Power Supply 7 Year Redux
Many people ask about long term computer power supply testing, and simple truth is that it is too expensive for HardOCP to do in-house as it would require hundreds of thousands of dollars' worth of resources. However, we can give the inquisitive a non-scientific look at how well a personal PSU does in our testing 7 years later.
 

GiGaBiTe

2[H]4U
Joined
Apr 26, 2013
Messages
2,398
I am a longtime member of the website Audiokarma.org which largely focuses on vintage audio 30-60 years old. All of my 40 year old gear HAS indeed been gone over by qualified technicians, and aside from some cracked solder joints, not found to have any issues with their capacitors.

I've seen lots of "qualified" audio gear techs over the years and rarely do they have a proper capacitor tester which tests capacitors up to their rated working voltage. Even expensive hand held LCR meters only apply 5-12v across the capacitor, which is not a complete test. If you test a 500v cap at 12v, it can work fine and dandy, but it doesn't show that breakdown could be happening at something like 50v, or any range up to 500v. The only people I know with such equipment are usually radio operators.

II at no point said that there can't be age related failures. What I said is "Age itself isn't really the biggest variable", which is true.

Age is the absolute biggest variable. Unless you're the original owner of the equipment, you have no way of knowing the history of the device.

To be concerned about age with these after only 10 years is crazy.

Not sure if you read data sheets, but capacitors of the electrolytic kind have rated lifespans of between 2000-10,000 hours at their rated temperatures (usually 85-105c.) This means in as little over a year the cap can have expired. And considering that in cramped switching power supplies where caps are often stuffed under hot heatsinks near smokin hot mosfets and in between hot switching transformers with little to no airflow around them, it's most definitely a concern in 10 year old power supplies with even the absolute best capacitors.

You can't compare wide open audio gear to a computer power supply. Good audio gear usually runs on linear power supplies, which doesn't smash caps with 20kHz+ frequencies like a computer switching supply does. They also tend to be much more open and not place capacitors right next to smokin' hot components.

"And how the fuck did you actually post that "It matters not if a capacitor was run smokin' hot, babied or never run at all"? Did you actually type that with a straight face? Temperature is absolutely the #1 thing that helps determine the life of most electronics, especially if we are talking decades.

Probably not the best worded sentence, but as per the second part of the sentence you ignored; Yes, time is the biggest determining factor regardless of the operating condition of the capacitor. If you run it smokin' hot, it's going to wear out more quickly. If you baby it, it's still going to wear out over time. If you don't use it at all, it's going to die due to electrolyte decomposition, which can go acidic/basic and damage the internal structure of the cap. The rubber plug can also dry-rot and allow air/electrolyte leakage and also cause failure over time.
 

Susquehannock

Supreme [H]ardness
Joined
Jul 26, 2005
Messages
4,401
PC Power & Cooling 425w unit bought march 31, 2004 for $140 running in WinXP rig constantly. Tested under load earlier this year and still within 2% or so on all three rails. Same main board, GPU card, and RAM used throughout. Clean input from industrial UPS is no doubt responsible for longevity of system as a whole.
 

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
54,075
I have two 1200w PSUs that sitting on my test bench that I think we will run back through testing this year. Paul has another that I think is 9yo that is about to be retired that we will likely to the same with.
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
11,358
I've seen lots of "qualified" audio gear techs over the years and rarely do they have a proper capacitor tester which tests capacitors up to their rated working voltage. Even expensive hand held LCR meters only apply 5-12v across the capacitor, which is not a complete test. If you test a 500v cap at 12v, it can work fine and dandy, but it doesn't show that breakdown could be happening at something like 50v, or any range up to 500v. The only people I know with such equipment are usually radio operators.

The tech I took it to has a great reputation and has fully refurbished many units (hundreds) including upgrading capacitors, I have no reason to doubt his abilities, and neither do others who paid him much more than I did.

Age is the absolute biggest variable.

I still believe heat is the #1 concern, but if you are not negligent, and actually maintain good airflow and/or other cooling, then something else would become the variable, and yes that could simply be age. I still can't conceive a situation where a properly cooled power-supply is going to fail in 10 years or less.

Unless you're the original owner of the equipment, you have no way of knowing the history of the device.

Absolutely. If you are buying used power supplies from people you don't know, that could be an issue. They could have never been dusted out and simply roasted internally for years. OP said his father bought them, so I think the history is pretty well known.

Not sure if you read data sheets, but capacitors of the electrolytic kind have rated lifespans of between 2000-10,000 hours at their rated temperatures (usually 85-105c.)

85-105c is pretty hot... Guaranteed none of my 10 year old computer powersupplies run that hot or ever have.

This means in as little over a year the cap can have expired. And considering that in cramped switching power supplies where caps are often stuffed under hot heatsinks near smokin hot mosfets and in between hot switching transformers with little to no airflow around them

Agreed 100% that heat kills. Thankfully my 10 year old PSUs have powerful fans that give great airflow to every part of the unit internally. Big fans were already very common in PSUs 10 years ago.
 
Last edited:

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,600
Wouldn't worry about it if it's a reliable brand.

I still have a 12 year old Silverstone Zeus 650W PSU in my secondary PC, it outlived every other component from the original build in 2005.

I'm still running the Zeus in my sig that will be 11yrs old in a months time, torn apart and benched by the man JG himself. :D

It has been used and abused through 4 different builds. All hail Etasis!

http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=5
 

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,600
Nice!

I thought I was just lucky but it seems like Silverstone is truly a reliable brand.

Like all PSU's, it depends on who makes the guts of the unit. Not all models from a manufacturer are created equal.
Our particular Zeus line was made by Etasis, which was known for bullet proof server PSU's.

Silverstone does make good PSU's, but sometimes their OEM of choice is questionable on certain models.
 

EchoWars

Limp Gawd
Joined
Jan 7, 2010
Messages
497
Yes, time is the biggest determining factor regardless of the operating condition of the capacitor.
This is the shit I do for a living...heat is the #1 enemy, not time (though time is a factor, provided that we're starting with a decent quality cap in the first place).

Not sure if you read data sheets, but capacitors of the electrolytic kind have rated lifespans of between 2000-10,000 hours at their rated temperatures (usually 85-105c.)
Can't think of too many environments (make that zero) in a home where a cap might be subjected to a temp anywhere near that. I've got the formula for calculating expected lifespan at temperatures under the rated max, and lifespan rises exponentially. 2000 hours @ 105°C goes to 50,000+ hours @ 35°C. You also mentioned something about the switching frequency...the cap don't care. Falls under the self-heating part of the ripple current rating.

I still believe heat is the #1 concern,
And you would be correct. (and if you're a member at AK, you might have seen my name pop up a couple of times :LOL:)

85-105c is pretty hot... Guaranteed none of my 10 year old computer powersupplies run that hot or ever have.
Ditto.

My biggest concern about a 10 year-old PS is: was it built with decent quality caps (Panasonic or Nichicon or Nippon Chemi-Con)? If it has cheap-ass Chinese caps in it, I'd personally be recapping it with something decent for peace of mind. If that's not your thing, and you're not up to digging inside your PS, shell out the $$ for a nice new one.
 
Last edited:

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,600
This is the shit I do for a living...heat is the #1 enemy, not time (though time is a factor, provided that we're starting with a decent quality cap in the first place).

If I had a nickel for every time I've told my clients this....
Which is why I steer them all away from anything passive, and have plenty of airflow...heat is your worst enemy when it comes to electronics.
 

manny1222

Gawd
Joined
Mar 11, 2012
Messages
732
I don't mean to hijack your thread but i was about to post something similar. I have a HX1050 that i bought used (i think, or refurbished). I think i bought it in 2012 or 2013 because i was doing gtx 680 sli. I just got a 1080ti and I'm planning to get an 8700K when its released.
Is it time for a new psu? I havent had any issues with the this one but i dont know what voltage numbers look like.
 

acidrain97

Limp Gawd
Joined
Nov 16, 2007
Messages
217
I still have a Soyo amd 300mhz with some crappy generic AT power supply. Put a new quiet 80mm fan and works great. Sure it's a pure dos box but who cares. I like my dos games played on real hardware. The PS is from 1994 or so. it's been in a few machines. I could install an atx (the board can handle AT or ATX) but why bother.
 

manny1222

Gawd
Joined
Mar 11, 2012
Messages
732
Is there a way test the psu, that an average guy like me that is not Kyle or Guru3D can do?
Should i do a cpu and gpu stress and monitor voltages in HWInfo64 or something like that?
 

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,600
Is there a way test the psu, that an average guy like me that is not Kyle or Guru3D can do?
Should i do a cpu and gpu stress and monitor voltages in HWInfo64 or something like that?

A multimeter will work, or if your not comfortable with that, you can purchase a PSU tester.
 

EchoWars

Limp Gawd
Joined
Jan 7, 2010
Messages
497
Those cheap 'Go/no-go' testers are pretty much a waste of time. Unless you have the ability to load the supply, yer just spinnin' yer wheels.
 

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,600
Those cheap 'Go/no-go' testers are pretty much a waste of time. Unless you have the ability to load the supply, yer just spinnin' yer wheels.

Very true, same goes for testing with a multimeter with no load, that's why I said use a multimeter first, you would be amazed on how many people are afraid of using them.
I should of probably mentioned that the testers are really only good for a non booting Pc.
 

EchoWars

Limp Gawd
Joined
Jan 7, 2010
Messages
497
Recommendation from a old tech: If you don't have any confidence in what you're doing, put the supply in and if everything runs, call it 'good' and get on with your life. Seen too much chit blowed up from misplaced probes or funky $10 Harbor Freight meters.
 

Hagrid

[H]F Junkie
Joined
Nov 23, 2006
Messages
9,157
Recommendation from a old tech: If you don't have any confidence in what you're doing, put the supply in and if everything runs, call it 'good' and get on with your life. Seen too much chit blowed up from misplaced probes or funky $10 Harbor Freight meters.
Or just buy a new PS. They are not expensive and would give the OP less worry.
When you buy a new car do you want used parts in it?
It`s only money!
 

WhoBeDaPlaya

2[H]4U
Joined
Dec 16, 2002
Messages
2,571
Not unless you're pushing it to 75%+ of its max rated load (or thereabouts).

I'm still using a ton of BFG ES-800 and modded EX-1000s that are easily 7-8 years old at this point.
Hell, I just put a CoolerMaster Real Power Pro 850 back into service after it was in storage for 5 years.
 
Last edited:

RFOneWatt

n00b
Joined
Mar 6, 2014
Messages
30
We can't be talking about the age of a power supply in YEARS. Just too many variables. We need to be addressing the age in HOURS.

With that said, the next two questions are:

What kind of an environment has the PSU been living in?
At how much of its rated capacity has it been running?

A high quality power supply that has been run at around 50-60%~ of its rated capacity in a temperature stable environment will last a helluva long time..

With that said, when in the market for a power supply it's a good idea to buy something a bit larger than what you "need."

Just my 2 cents.

~RF
 
Top