How to get caps that wont fail?

[chico]

I was wondering how you can get caps that wont fail.
Capacitors seem to be the main component to fail in any electronics that I fix.
What causes this?
Current that's too high?
Temperature effects?

The hubbel telescope has caps im sure and it works for many years, how do I avoid buying caps that fail?

 

[Nigel Goodwin]

Don't buy cheap crappy ones - it's those that fail, the most common offenders are ones made by Capxon.

Buy good quality Panasonic or Rubicon, low ESR and 105 degree, and you shouldn't have any problems.

If you check the specs on the Capxon ones that fail all the time, they are only rated for about 15 months life at the temperatures they are usually used in.

 

[cobra1]

the hubble telescope probably has caps in it worth $1,000,000 lol

 

[tvtech]

Don't buy cheap crappy ones - it's those that fail, the most common offenders are ones made by Capxon.

Buy good quality Panasonic or Rubicon, low ESR and 105 degree, and you shouldn't have any problems.

If you check the specs on the Capxon ones that fail all the time, they are only rated for about 15 months life at the temperatures they are usually used in.

Some wise advise:

Electrolytic caps fail mostly in smoothing Switch Mode PSU Secondary rails. You will not easily find a Main Smoothing cap that has failed on the Primary side (rail) due to issues other than overvoltage..(Mains too high). 85 Degree caps are good for the Mains Smoothing side. And very common. No matter the manufacturer.

The Secondary rails are another issue however. These caps have to deal with smoothing higher frequency switching. And it is a vicious circle for an inferior cap...

Chemistry is bad to start with. So the cap heats up to cope with the required smoothing. The more it heats up, the more it looses smoothing capacity. The more it looses smoothing capacity...the hotter it gets and so on. And pops or bursts That is the simple explanation.

Believe me, if a Chinese cap says on it's casing it is a 105 Degree cap...be very wary.

Now back on topic:

The same junk caps used in low ripple or unstressed environments will work for ages. No heat and no stress = long life. Like a portable radio.
Who of you Gurus have ever replaced an Electrolytic cap on a Battery powered radio and it fixed the problem??

Cheers,
TV TECH

 

[bountyhunter]

I was wondering how you can get caps that wont fail.
Capacitors seem to be the main component to fail in any electronics that I fix.
What causes this?
Current that's too high?
Temperature effects?

The hubbel telescope has caps im sure and it works for many years, how do I avoid buying caps that fail?

What is the application? In general, you get what you pay for. Aluminum electrolytics give most capacitance for $, and have the shortest operating life. Tantalums have better life if used properly. Ceramics last a long time, but are costly for capacitance delivered. Each type cap has it's advantages and disadvantages.

FYI, the reason caps fail so much in consumer products is the manufacturers go cheap on them to save cost.

 

[Nigel Goodwin]

Electrolytic caps fail mostly in smoothing Switch Mode PSU Secondary rails. You will not easily find a Main Smoothing cap that has failed on the Primary side (rail) due to issues other than overvoltage..(Mains too high).

Actually failure is reasonably common, and not just through overvoltage (which is EXTREMELY rare in the UK - but very common in Spain!) - usual faults are the electrolytic O/C, and you can often see where the electrolyte has leaked out.

 

[bountyhunter]

Electrolytic caps have MTBF numbers that are orders of magnitude less than IC's. many are in the thousands of hours, and it drops dramatically as temperature increases. The reason switcher stress caps is because the high frequency ripple current flowing through the ESR causes internal heating which shortens the life. You can improve it by using lower ESR caps (more $$) or paralleling more caps (more $$) or going to a better type cap (more $$)

 

[OlPhart]

I'll second & give bonus points to bountyhunter for a simple, succinct answer... <<<)))

 

[MrAl]

I was wondering how you can get caps that wont fail.
Capacitors seem to be the main component to fail in any electronics that I fix.
What causes this?
Current that's too high?
Temperature effects?

The hubbel telescope has caps im sure and it works for many years, how do I avoid buying caps that fail?

Hi,

It's both the current and the temperature that have major effects on the life of the capacitor.

Electrolytic caps have lifetimes in thousands of hours at a given temperature, such as 1000hr at 85 degrees C. There are about 9000 hours in one year.

The saving grace is that capacitor life approximately doubles for every 10 degree C decrease in temperature, so that same cap is good for 2000hr at 75 deg C, 4000hr at 65 deg C, 8000hr at 55 deg C, etc. However, a cap rated at 1000hr at 105 deg C will live for 2000hr at 95 C, 4000hr at 85 C, 8000hr at 75 deg C, 16000hr at 65 deg C, 32000 hr at 55 deg C (about three or four years).
It's plain to see that the higher the initial life rating in hours and the higher the initial temperature rating the longer the cap will last.

Ripple current plays a role here too though because high ripple makes the electrolyte dry out faster. Polymer caps are supposed to be better for this though, although the proof is in the life vs temperature rating again.

You can find some caps rated for 2000 hours at 85 deg C and some rated at 10000 hours. Obviously choose the higher rating. Of course the price will reflect the better quality too in most cases.

I had a computer power supply that would shut the computer down every so often without warning. Took it apart, found several electrolytic caps that leaked. Their capacitance went way down and their ESR went up quite a bit. Replaced all of them. Cost about 10 dollars or so to get the better grade caps in there. Will probably last at least another 5 years.
Actual cost to buy the caps was a little higher than that though because i couldnt buy just 1 of each cap i needed, and i wanted a few extra anyway. Also, one cap size could not be matched perfectly so i ended up with a slightly larger diameter cap which is better for life (same voltage rating) so the other caps had to be mounted more carefully because it took up slightly more room.

 

[bountyhunter]

I had a computer power supply that would shut the computer down every so often without warning. Took it apart, found several electrolytic caps that leaked. Their capacitance went way down and their ESR went up quite a bit. Replaced all of them. Cost about 10 dollars or so to get the better grade caps in there. Will probably last at least another 5 years.
Actual cost to buy the caps was a little higher than that though because i couldnt buy just 1 of each cap i needed, and i wanted a few extra anyway. Also, one cap size could not be matched perfectly so i ended up with a slightly larger diameter cap which is better for life (same voltage rating) so the other caps had to be mounted more carefully because it took up slightly more room.

When I looked for replacement PC power supplies, I found new ones as cheap as $12. Here is one for $14.

**broken link removed**

If anybody wonders why the PC power supplies fail, just look at the price. You get a 400 watt power supply complete and tested for under $20. There is absolutely no profit margin in them. They have to go cheap on every single part they can.

 

[MrAl]

Hi,


I've also read that there have been an influx of defective Chinese electrolytic caps that blow out much faster than expected. I suspect this because the power supply wasnt that old really.

 

[Sceadwian]

14 dollars? That sounds way more than too good to be true.. the scrap value of the metal in that is probably close to that price... I wouldn't touch something that cheap with a 10 foot pole literally!

 

[bountyhunter]

If you buy any PC you are touching them. They all use the same power supplies made by the cheapest bidder on the Pacific Rim. The PC makers probably pay less than $10 for them.

 

[Sceadwian]

Hell no Boutyhunter, not a chance. The supply in my machine is of SIGNIFICANTLY higher quality than that piece of junk. You just have to watch out for brand names.

 

[bountyhunter]

Hell no Boutyhunter, not a chance. The supply in my machine is of SIGNIFICANTLY higher quality than that piece of junk. You just have to watch out for brand names.

Maybe. I just bought a brand newt top end Hewlett Packard PC to replace the emachines that blew up (power supply failed BTW) and the power supply in the new HP failed in three days. They sent me a replacement that has been working OK for all of two months so far. They are all made in China and they are all junk.

 

[Sceadwian]

That's because Hewlett Packard makes junk computers.. I have never once in my entire life (Though contact with a dozen HP machines) every had a positive experience with HP products.

Dude, you should have gotten a Dell New Dell, not old Dell. We have two machines through them and they're rock solid, I've worked on a few others, never had a problem, I clean ours regularly, everything from case design aesthetics internal air flow patterns to ease of access case tabs are all top notch.

Not to say that parts of our machines aren't made in China, that's just the way things are, but it isn't a throw away black label CE/UL branded slave labor hardware.

I won't outgrow this machine for at least another 2-3 years, and it's 4 years old. Thermal paste is probably due for a change.

It's just like a car, if you take care of it, it'll keep on going. I can thank this to the depreciation rates of PC's, it was a 1200 dollar (US$) computer, I bought it three months after the owner needed to sell it (transferable warranty) for 600 and that was fair.

 

[bountyhunter]

Glad you got a good one. You do realize what "anecdotal" means..... your experience is far from universal.

**broken link removed**

**broken link removed**

http://answers.yahoo.com/question/index?qid=20090427124432AAubIRC

http://dell-computer-reviews.measuredup.com/Complaint-Dell-Computer-reliability-11264

IMHO, PC hardware is all junk.

 

[Sceadwian]

Based on personal observation, case study reports, or random investigations rather than systematic scientific evaluation: anecdotal evidence.

I research what I buy before I throw cash on the table extensively. Hasn't failed me yet.

 

[ronv]

Several years ago I worked for a disk drive manufacture. Dell was our largest customer and we sent them the best stuff. What I'm saying is if we had 2 sources for heads and disks and one was better than the other Dell got the best ones. We were building disk drives for about $30 complete - probably less now. I ran the failure analysis department and we analyzed thousands of drives. I'm not sure I can remember a bad cap, but there weren't many and they were all in low stress circuits.
I agree with the other posts on caps: Temperature, Ripple current, ESR and temperature rating.

 

[bountyhunter]

I research what I buy before I throw cash on the table extensively. Hasn't failed me yet.

I believe you got a good one. My point is all PC makers have a ton of angry customers on the net. Their quality control is crap. It is the inevitable and absolutely predictable end of any product market that does the death spiral down on price.

When PCs first came out, it was easy to crank up price by differentiating performance... ie, you need the new whiz bang processor that is TWICE AS FAST... or the new computer with TWICE AS MUCH MEMORY..... the problem is that CPU performance is so fast for the buck and memory capacity is virtually unlimited for cheap $, so the vast majority of customers won't pay more money for performance differentiation. Once that happens in any product, the only "give" is in the price hence the cost of PC's goes down. Good in one respect, bad in others: as cost gets cut to the bone quality control goes away because they can't afford it.