When the PSU is weaker and dying...

[WaiWai]

When the PSU is weaker and dying...
Aging is the problem of every component. This will affect the power supply.

1) How serious is aging? How much percentage of power supply will be lost after, say, 5 years of usage? Is there any article which research the aging effect?

2) If my PSU starts to be weaker or start dying, it may not be able to provide enough power or do other weird things. I, as an end-user, can't test/check it periodically.
As an end-user point of view, how can I detect the possible failure before the PSU can take the chance to damage my other hardware?

Thanks.

 

[Zap]

1) Pretty serious problem, but whether it affects you depends on how long you keep power supplies. If you tend to upgrade your PSU (or replace whole system) every year or two, you probably will never encounter the problem. If you are still hanging on to your old Celeron 300A system with all original parts, you might be concerned.

Don't know about any articles on capacitor aging, but there is a lot of evidence and discussion in recent years based on product failures due to aged capacitors, such as all those motherboard leaky caps from a few years back, or leaky caps in original Antec TruePower (and other) PSUs.

2) You might be able to detect it using a multimeter or oscilliscope while the system is under load. Just do a test when the PSU is new, and then just test it periodically. If the numbers start to stray, then you know something is wrong.

 

[beray]

Originally posted by: WaiWai
When the PSU is weaker and dying...
Aging is the problem of every component. This will affect the power supply.

1) How serious is aging? How much percentage of power supply will be lost after, say, 5 years of usage? Is there any article which research the aging effect?

2) If my PSU starts to be weaker or start dying, it may not be able to provide enough power or do other weird things. Is it possible to detect the possible failure before the PSU would damage my other hardware?

Thanks.

---- How serious is aging?

Not serious at all if you know what you're doing. Electrolytic caps at room temperature can last more than 20 years. eg, don't let your PSU get hot if wanted a long life for your PSU.

---- How much percentage of power supply will be lost after, say, 5 years of usage?

You can call it lost and in a way it is, but usually malfunctions show up first.

If you don't let your PSU get hot, you can ignore 5 years of usage completely. If your PSU had been blowing out heat like an oven then you won't have to worry, it would die before 5 years is up.

---- Is there any article which research the aging effect?

Don't know of any now but the basic rule of thumb is --> For every 10 degree Celsius increment, the cap life is shorten by half.

---- Is it possible to detect the possible failure before the PSU would damage my other hardware?

Definitely so.

First you need the second rule of thumb --> The cap life is directly proportional to its size. The smaller the caps the quicker they became desiccated and lose capacitance.

With that second rule you ignore the PSU input storage caps, and check for the aging of output filter capacitors.

What you are looking for is the magnitude of output noise ripples at worst case loaded and unloaded, if the ripples are twice the normal size or exceeding standard specs on any power rail, replace all electrolytic caps including the input storage caps then enjoy your brand new refurbished PSU again.

PS. The above do not include coverage for PSU design weaknesses nor defective parts.

 

[Zap]

Originally posted by: beray
don't let your PSU get hot if wanted a long life for your PSU.
...
PS. The above do not include coverage for PSU design weaknesses nor defective parts.

I like your PS.

As for PSU temps, thanks for pointing that out.

 

[WaiWai]

Originally posted by: Zap
1) Pretty serious problem, but whether it affects you depends on how long you keep power supplies. If you tend to upgrade your PSU (or replace whole system) every year or two, you probably will never encounter the problem. If you are still hanging on to your old Celeron 300A system with all original parts, you might be concerned.

Don't know about any articles on capacitor aging, but there is a lot of evidence and discussion in recent years based on product failures due to aged capacitors, such as all those motherboard leaky caps from a few years back, or leaky caps in original Antec TruePower (and other) PSUs.

I hope to use 3-6 (more aggressively, 6-9) years.

I will pick a decent/good brand.

It happens a good PSU has a lifespan more than 5 years on average.

I'm interested to know how much power it can still supply after 5 years, for example. If it doesn't drop too much (eg 40-50%), I'm interested in buying a slightly more powerful PSU. For example, if I know my PSU can actually supply 400W which is more than enough for a max 300W system. I may buy a PSU which can actually supply 500W (accounting aging issues) so I can save 1 upgrade of PSU when I upgrade my PC in about 3 years.

2) You might be able to detect it using a multimeter or oscilliscope while the system is under load. Just do a test when the PSU is new, and then just test it periodically. If the numbers start to stray, then you know something is wrong.

I will only do if it's proved to be the only effective way to do. I prefer other ways.
I don't think I will have the time to keep checking it periodically. I am going to fail due to laziness or busy working.

 

[beray]

@Wai Wai

Here you go http://www.gigabyte.com.tw/Fil...ticle_02_all_solid.htm

 

[WaiWai]

Not serious at all if you know what you're doing. Electrolytic caps at room temperature can last more than 20 years. eg, don't let your PSU get hot if wanted a long life for your PSU.

Good to hear because I'm going to pick a decent-to-good PSU. Their capacitors should be much more resistant to aging.

You can call it lost and in a way it is, but usually malfunctions show up first.

If you don't let your PSU get hot, you can ignore 5 years of usage completely. If your PSU had been blowing out heat like an oven then you won't have to worry, it would die before 5 years is up.

Sorry for a bad question.
What I mean is approximately how much the PSU can still supply after 5 years.
For example, a PSU can supply 500W at the start of Year 0 but 250W at the end of Year 5. The power loss is 50%.

What you are looking for is the magnitude of output noise ripples at worst case loaded and unloaded, if the ripples are twice the normal size or exceeding standard specs on any power rail, replace all electrolytic caps including the input storage caps then enjoy your brand new refurbished PSU again.

How can I monitor the output of noise ripples?

I am unable to keep testing/checking my PSU periodically. I hope there is a convenient way, as an end-user point of view, to detect it so I can prevent my PSU from damaging my hardware.

 

[beray]

Originally posted by: WaiWai

Sorry for a bad question.
What I mean is approximately how much the PSU can still supply after 5 years.
For example, a PSU can supply 500W at the start of Year 0 but 250W at the end of Year 5. The power loss is 50%.

There is no loss, the first electrolytic caps to be affected by aging are the smaller ones used for output filtering. The higher noise ripple levels would cause problems way before anything else became noticeable.

When the PSU input storage capacitors are affected by aging, you'll have an unstable power supply because the input power holdup time is too short.

Originally posted by: WaiWai

How can I monitor the output of noise ripples?

I am unable to keep testing/checking my PSU periodically. I hope there is a convenient way, as an end-user point of view, to detect it so I can prevent my PSU from damaging my hardware.

I don't know of any practical methods which would allow the monitoring of high frequency noise ripples other than 100kHz-voltmeters, Scopemeters, etc...