[SOLVED] Old PSU performance degradation tests?

[Rinshun]

Hello there

Are there any tests comparing the performance of new and used PSUs?

We all know the power supplies will degrade over time, but at what rate and which aspects will degrade faster? Ripple? Transient response? Power output?

 

[kanewolf]

Hello there

Are there any tests comparing the performance of new and used PSUs?

We all know the power supplies will degrade over time, but at what rate and which aspects will degrade faster? Ripple? Transient response? Power output?

What test equipment do you have? Multi channel O-scope? Dedicated load bank? Multple DVMs?

 

[Grobe]

We all know the power supplies will degrade over time, but at what rate and which aspects will degrade faster? Ripple? Transient response? Power output?

Well, considering the electrolytes are the main component type that is actually capable of degrading over time, expect multiple problems, including controller circuit probably have to overcompensate all the time is itself a problem that may lead to unstability characteristics that I don't know the name of.

 

[Rinshun]

What test equipment do you have? Multi channel O-scope? Dedicated load bank? Multple DVMs?

I have a microwave and an electric toaster

 

[Karadjgne]

The only way to test is by comparison. You'd need to to have brand new components and push them to the breaking point. With a large enough sample base to decrease margin of error, then test the components of the psu in question and compare the results. Does it hold the same charge, has capacitance changed, are voltages and current %'s within error etc and push it to the same breaking point of the control sample, does it hold, is the ripple worse etc.

Corsair has done it, Seasonic, SuperFlower and prolly several other companies, it's part of how they come up with their warranty period. They know from experience and design just what their psu's require in component choices, why use Rubicon caps over similar rated CapXon etc.

Apart from the tools mentioned above, you'd also need a competent education in electronics engineering and an intimate understanding of the relationship between component quality and lifespan.

As to what rate? Which is faster? That's totally dependent on the psu itself, usage, age, abuse, maintenance etc. It's one thing to test psus in a clean room with simulated loads, it's entirely different story to grandma's Dell web browser she's had for 10 years and never once been dusted or cleaned.

 

[Rinshun]

The only way to test is by comparison. You'd need to to have brand new components and push them to the breaking point. With a large enough sample base to decrease margin of error, then test the components of the psu in question and compare the results. Does it hold the same charge, has capacitance changed, are voltages and current %'s within error etc and push it to the same breaking point of the control sample, does it hold, is the ripple worse etc.

Corsair has done it, Seasonic, SuperFlower and prolly several other companies, it's part of how they come up with their warranty period. They know from experience and design just what their psu's require in component choices, why use Rubicon caps over similar rated CapXon etc.

Apart from the tools mentioned above, you'd also need a competent education in electronics engineering and an intimate understanding of the relationship between component quality and lifespan.

As to what rate? Which is faster? That's totally dependent on the psu itself, usage, age, abuse, maintenance etc. It's one thing to test psus in a clean room with simulated loads, it's entirely different story to grandma's Dell web browser she's had for 10 years and never once been dusted or cleaned.

So... any independent reviewers did this? Like GamerNexus, etc?

 

[kanewolf]

So... any independent reviewers did this? Like GamerNexus, etc?

A reliable site will post the tools and processes they use to test power supplies. Reviewers can't do large samples. But they will typically send anomalies to the manufacturer for comment and analysis.

 

[Karadjgne]

I don't know as I've ever seen Any reviewer, anywhere at any time do any sort of degradation testing. Basically, they get a brand new sample, either store bought or factory cherry picked, run it through their particular test procedures, and maybe tear into it to see what components are actually used. After that it goes on the shelf, goes in a build, gets donated or dumpstered.

And then on to the next project. Most won't use the samples for working builds, maybe one or two excellent choices, and certainly won't keep the builds working for 5-10 years in order to actually get a working sample to then degrade test.

There's plenty of actual proof of caps degrading, motherboards are common enough, but actual testing to see how much, doesn't really make good headlines and is somewhat cost and time prohibitive.