General PSU Cap Failures

[Roving Geek]

Hello All. I hope that i have not overlooked information that has been previously posted, and if so, I apologize for my impatience or lack of diligence in searching more thoroughly.

I have several failed PC ATX PSU's from varied manufacturers, and out of curiosity, I finally opened them up to examine them.

I have noticed that in a majority of them, there is a 2200 uF 10v cap and a 1000 uF 10v cap that are bulged. The 2200uF cap is slightly larger in diameter than the 1000uF cap.

As far as I can tell, the schematic locations of these caps are common to the failed PSU's.

The PSU's all have the 4-pin 12v connector for that connects to the motherboard near the CPU socket, and the power connections are a even mix of 20 pin and 20+4 pin configurations.

My first question, what specifications are recommended for the replacement caps?

And my second question is, what may be the common cause in these two caps failing? I believe that the units all came form PC's that were not protected buy a battery UPS unit, or had very cheap power strips.

[Per Hansson]

Re: General PSU Cap Failures

My first question, what specifications are recommended for the replacement caps?

Specifications that are close to the originals.
And capacitors manufactured by one of the major Japanese brands.

And my second question is, what may be the common cause in these two caps failing? I believe that the units all came form PC's that were not protected buy a battery UPS unit, or had very cheap power strips.

The common thing they will have is that they are made by cheap ass Chinese capacitor brands that can do only one thing:

[SteveNielsen]

Re: General PSU Cap Failures

https://www.badcaps.net/pages.php?vid=4

[Roving Geek]

Re: General PSU Cap Failures

The common thing they will have is that they are made by cheap ass Chinese capacitor brands that can do only one thing:

Thank you for the input, Per Hansson, but I was hoping for a bit more. I already came to the conclusion that low quality material was a contributing factor. But I curious as to whether poor line voltage had any significance.

I have always used UPS units on my computers. I am not much into totally geeking out and building the ultimate computer, so I have leaned more to what has been typically off-the-shelf units. My personal units seemed to have been spared from PSU failures or MB failures, unlike a lot of the computers that I come across to be repaired or serviced.

[Per Hansson]

Re: General PSU Cap Failures

If you want an answer that is more to the point we are obviously going to need more info then: I have a car, it's tires are worn, why do tires get worn?

[Roving Geek]

Re: General PSU Cap Failures

Touche! General information or questions generally produce little, as you so pointedly stated.

I acknowledge I framed my inquiry poorly. Therefore, allow me to restate my inquiry.

I have observed that failed computer power supply units in consumer class computers appear to experience 2 capacitors that fail. They are either a 2200uF 10v and a 1000 uF 10v, or 2 2200uF 10v. These caps appear to me to be schematically commonly located, regardless of the brand.

Now, given that the caps are from poor quality suppliers, would these failures be less likely if adequate line voltage protection been used?

https://www.badcaps.net/pages.php?vid=4

SteveNielsen's post is very informative but seems to lay the blame on the poor quality of the capacitors used during manufacture.

If the components are of such poor quality, how could they get pass even the minimal testing the manufacturers subject their products to in order to get approved by the major computer manufacturers (HP, Dell, Lenovo, etc)?

[Per Hansson]

Re: General PSU Cap Failures

...would these failures be less likely if adequate line voltage protection been used?

No, the output capacitors are not effected by the incoming power.
Only the primary side capacitors are affected by poor line quality.
(Well it's an oversimplification obviously but still a good general rule)

SteveNielsen's post is very informative but seems to lay the blame on the poor quality of the capacitors used during manufacture.

If the components are of such poor quality, how could they get pass even the minimal testing the manufacturers subject their products to in order to get approved by the major computer manufacturers (HP, Dell, Lenovo, etc)?

The blame is on the capacitors, it can also be a poorly designed PSU in general, which puts more stress on the capacitors.
But in general these things go hand in hand.
Meaning poor cheap ass Chinese crap PSU is paired with capacitors of likewise quality.

While good quality power supplies also get good quality capacitors, when in effect it's the cheap ass PSU that would benefit the most from it.
(Not that I'm condoning the use of cheap ass caps in quality powersupplies, in case you are reading this Corsair!)

As for your question about testing I'm sure the Chinese are able to produce a fake testing report, failing that a decent PSU and after the test they take half of the components out.
Or replace them by cement, like in some fine quality powersupplies from China!

[mariushm]

Re: General PSU Cap Failures

The capacitors you mention are on the secondary side of the power supply, so the amount of line protection or anything related to mains is irrelevant.

One of the capacitors you mention (rated for up to 10v) is used by the 5v stand-by circuit, which is sort of like a mini power supply INSIDE the power supply. This "mini power supply" is running 24/7 even when your computer is turned off. Therefore, it's quite normal to see that capacitor fail more often than other capacitors in the power supply.

As for the others, and for failure of capacitors in power supply in general..

Due to way these switching power supplies work, those capacitors have to be physically located as close as possible to other components in the power supply, the inductors and some diodes or mosfets.
Unfortunately, those components are also generating a lot of heat... even the best power supplies out there are only 92-94% efficient, so all the rest is basically heat which has to be moved outside the case.

Now, there's all kinds of capacitors but the kind of capacitors needed for switching power supplies are the "low esr" type (or low impedance).
There's all kinds of chemicals and solutions (formulas) used to produce the electrolyte in capacitors, capacitors use different formulas depending on what they're designed for.

The more "high performance" a capacitor is, in general the substances used inside them are more sensitive to heat.
Some brands of capacitors tolerate heat better, some less... hence why you see the capacitors fail. A premium power supply will not only use higher end capacitors (more expensive) but will also design the circuit better (for example my Seasonic X-650 has some heat generating components on the back of the PCB connected directly to the power supply case metal, which acts like a huge heatsink) and will have better quality fans and maybe the cables inside are also routed better to make the cooling better.


How well the capacitors and the other parts are kept cool in the power supply depend on lots of things.

Customers demand low price, so manufacturer uses smaller heatsinks to reduce weight and use less expensive capacitors reasoning that the power supply has only 1-2 years warranty and pretty much they'd last double that so they're in the clear.

Smaller heatsinks would be fine if there was more air cooling, but customers also demand silent power supplies and therefore you have power supplies that dynamically adjust the fan speed, potentially causing overheating during some periods.

Other things that you have to keep in mind is dust, how fast the air vents are blocked or clogged by dust, by hair, by cigarette smoke.. manufacturer can't predict all that.

So what else can I say.. it's incorrect to say the components are of such poor quality they're the reason power supplies fail, you have to look at the whole to see the compromises made and blame customers that buy the cheapest stuff forcing companies to make compromises.

[Roving Geek]

Re: General PSU Cap Failures

Thank you. Now I think I understand more. I agree that pencil-whipped documentation would naturally go hand-in-hand with manufacturing methods devised to maximize profit at the cost of a reliable product. Make items cheap enough so that it is simply more cost effective to just replace rather than service does seem to be the mantra of Chinese electronics.

And, Chinese manufactured power supply units is the very center of what I am looking at.

Again, I appreciate your patience and your logic in helping me to understand what it is that I observed.

Bottom line, if the unit is poorly designed or manufactured, in coming power regulation is futile.