Has anyone tested PSU before and after long term use?

[Kazualty]

Wondering if anyone has tested the long term effects of PSU use? Like if ripple increases, efficiency decreases etc. and by how much. Would be interested to know what happens to a power supply over time and which parts are at most risk of failing first (assuming you have a good quality PSU with no mechanical flaws). If anyone can point me towards a study or video or something would be appreciated, thanks!

[Tech87]

The Gamers Nexus channel have started PSU testing, and will be continuing their methodology in the future.

 

https://youtu.be/WToOFblsXqM

[Kazualty]

Just now, Tech87 said:

The Gamers Nexus channel have started PSU testing, and will be continuing their methodology in the future.

 

https://youtu.be/WToOFblsXqM

I don't just mean PSU testing but I have seen that and it's great to see GN expanding into PSU. 

 

I'm looking for long term effects of use. How efficiency and ripple might be effected over 5 or even 10 years. What kinds of degredation and how long can a PSU potentially last? All could make for an interesting GN video but I can't seem to find any information on it other than opinions. 

[Elisis]

2 hours ago, Kazualty said:

I'm looking for long term effects of use. How efficiency and ripple might be effected over 5 or even 10 years. What kinds of degredation and how long can a PSU potentially last? All could make for an interesting GN video but I can't seem to find any information on it other than opinions. 

I mean, efficiency and ripple obviously get worse and the extent to which this happens depends on the quality of the unit in question. Though I'd definitely be more concerned about the fan dying, say.

[Kazualty]

1 hour ago, Elisis said:

I mean, efficiency and ripple obviously get worse and the extent to which this happens depends on the quality of the unit in question. Though I'd definitely be more concerned about the fan dying, say.

Thats just logical assumptions as far as the ripple and efficiency but what's the amount for good PSU and what kind of life expectancy would you expect beyond the warranty.

 

Obviously you shouldn't risk beyond warranty on a new machine but if you are still using a 10 year old rig and keep no important information on your PC it might be worth risking for another couple years or maybe even another 10 but I don't have any actual information to base this on, just assumptions.

 

As far as fan that's pretty easy to replace for those willing to. I'd be more concerned about whether a capacitor is the first expected loss (beside fan) or something else? Which piece of hardware is most limiting and what's the actual losses over time for a good quality PSU. 

 

It will obviously vary from PSU to PSU and based on quality but would be nice to see even just one tested after long use and compared to previous tests. Of course I'd prefer more information than just one but I currently now have zero information to actually show how much degredation takes place. 

[Elisis]

5 minutes ago, Kazualty said:

I'd be more concerned about whether a capacitor is the first expected loss (beside fan) or something else? Which piece of hardware is most limiting and what's the actual losses over time for a good quality PSU. 

Not capacitors because capacitors don't produce heat, nor do they just straight-up die. I'd be more concerned about FETs, and perhaps magnetics because if those die, there can be huge consequences.

[f14]

I do it everyday. most of them not change so much (<2year of use). some lost power (example 500w psu and it shutdown when load 400W). and some blowup when you full load

ah all psu I test below 5 year of use ( because it cover by warranty time)

[Guest]

Whenever I've tested older units, the ripple is higher, the voltage is farther from nominal and they don't output as much maximum power as new.

 

[Kazualty]

2 hours ago, jonnyGURU said:

Whenever I've tested older units, the ripple is higher, the voltage is farther from nominal and they don't output as much maximum power as new.

 

Thanks for the responses and of course @jonnyGURUyou are a trusted source I believe that is what happens. I just wish I had some real numbers in front of me and a way to determine a longer lifespan.

 

If for example you don't have as much power being used and had a power supply with tight ripple to start with, would it then last significantly longer assuming the ripple isn't terrible and voltages are still within spec? Or if using a semi passive PSU but keeping the fan on at all times to extend life how much would this go beyond warranty assuming in passive mode it would have made it no problem?

 

I suppose that's why warranties are longer on some I just would like to see some real world results and comparisons. Would be nice to see a company like GN test a bunch of their older used PSU. Probably asking too much though.

[Guest]

9 hours ago, Kazualty said:

Thanks for the responses and of course @jonnyGURUyou are a trusted source I believe that is what happens. I just wish I had some real numbers in front of me and a way to determine a longer lifespan.

 

If for example you don't have as much power being used and had a power supply with tight ripple to start with, would it then last significantly longer assuming the ripple isn't terrible and voltages are still within spec? Or if using a semi passive PSU but keeping the fan on at all times to extend life how much would this go beyond warranty assuming in passive mode it would have made it no problem?

 

I suppose that's why warranties are longer on some I just would like to see some real world results and comparisons. Would be nice to see a company like GN test a bunch of their older used PSU. Probably asking too much though.

 

If the PSU has really tight ripple to begin with, it's very unlikely that it will be out of spec in 10 years time.  If it can no longer handle a load, it will likely just shut down.  I have seen some that don't, though.  Like, the MOSFET suffers thermal runaway and just blows up.  

 

As for warranty:

Length of warranty has more to do about marketing and the company's infrastructure and ability to support the product.  I've actually seen PSUs with 5 year warranties that are better than those with 10 and 12 year warranties.

[Kazualty]

On 12/1/2021 at 9:31 AM, jonnyGURU said:

 

If the PSU has really tight ripple to begin with, it's very unlikely that it will be out of spec in 10 years time.  If it can no longer handle a load, it will likely just shut down.  I have seen some that don't, though.  Like, the MOSFET suffers thermal runaway and just blows up.  

 

As for warranty:

Length of warranty has more to do about marketing and the company's infrastructure and ability to support the product.  I've actually seen PSUs with 5 year warranties that are better than those with 10 and 12 year warranties.

I appreciate the information. I kinda figured that was the case. I would still love to see someone do multiple tests on older good models of PSU to see how far they have degraded. Going from gold to bronze and losing even a couple hundred watts and even going from tight to not great ripple for most people probably wouldn't be terrible assuming the rest of the parts are still within life expectancy. 

[Guest]

26 minutes ago, Kazualty said:

I appreciate the information. I kinda figured that was the case. I would still love to see someone do multiple tests on older good models of PSU to see how far they have degraded. Going from gold to bronze and losing even a couple hundred watts and even going from tight to not great ripple for most people probably wouldn't be terrible assuming the rest of the parts are still within life expectancy. 

Unfortunately, there are so few reviewers that have the proper equipment to test PSUs,   The number of those with product old enough to go back and do retesting, assuming they still have old units, is even less.  It's a very, very, very specialized position to properly test PSUs.  I do it at Corsair so marketing and support know what to expect within the warranty period, but I don't expect every brand to do the same and make the information public.