Question about 85c caps vs. 105c caps

[PCBONEZ]

Re: Question about 85c caps vs. 105c caps

Do those particular 85c caps have any chance of lasting 5-7 years? Would a well ventilated case extend their longevity, judging by the temps I posted (if that helps at all)?

With good cooling they might last forever. Really hard to say.

Another issue is BGA vs. PGA. While this system is old enough to be of a PGA specification for the CPU socket, the motherboard (in Everest) came up as having a FCBGA specification. To my knowledge, Ball Grid Array is lead free and uses tin in lieu of lead whereas Pin Grid Array uses lead and has a better chance of (despite lower thermal resistance in theory) keeping the soldered connections to the board solid. Do you think that would have any chance of impacting the life of the board or the caps?

I'm not aware of RoHS being specific to an array type.
But neither type has been a problem with Intel CPUs AFAIK.
I wouldn't be concerned about it.
.

[PCBONEZ]

Re: Question about 85c caps vs. 105c caps

Those tests give an estimate, basically saying if the capacitors are subjected for various periods of times at temperatures close to that 85c or 105c, when the total amount of time exceeds that value, the capacitor will probably no longer be within the specified parameters (+/- 20% capacity and so on).

Kind of.
They put them at max temp, frequency, volts, and ripple current all at the same time for 'that' many hours.
They then cool them to 20C and measure them.
Everything except 1 [which I can't remember right now but I think it is DC Leakage] has to be inside the spec given in the data sheet.
.

[Wester547]

Re: Question about 85c caps vs. 105c caps

Thanks lots for the reply - very informative, again. (:

A few more questions...

Do those particular 85c caps have any chance of lasting 5-7 years? Would a well ventilated case extend their longevity, judging by the temps I posted (if that helps at all)?

Another issue is BGA vs. PGA. While this system is old enough to be of a PGA specification for the CPU socket, the motherboard (in Everest) came up as having a FCBGA specification. To my knowledge, Ball Grid Array is lead free and uses tin in lieu of lead whereas Pin Grid Array uses lead and has a better chance of (despite lower thermal resistance in theory) keeping the soldered connections to the board solid. Do you think that would have any chance of impacting the life of the board or the caps?

[PCBONEZ]

Re: Question about 85c caps vs. 105c caps

PCBONEZ, thank you for the response. I find it very informative.

Points of discussion I'd like to touch base on:

Is there an industry standard test performed on all caps/boards before they're put into production that affects the lifespan of all boards inevitably from day one (if I read correctly)?

The Endurance test is done to caps to get the required rating for the data sheet but not to those sold.

Chinese aluminum, are you sure no Japanese brands have Taiwan aluminum foil? I'm wondering about that in respect to KZG caps - I recollect that they're from Chemicon (Japan) yet they still die prematurely (after maybe 4-5 years of use). They don't have Taiwan aluminum?

Yes I'm sure. The Jap manufacturers all make their own foils, cans and electrolytes for QC purposes. They are in fact QC fanatics. It's a big deal to them. QC is a new concept in CH and TW and those that even try aren't very good at it yet. Many of them simply buy their foils, cans an electrolties so their quality depends on who they bought from last.
.
The one exception is the (Japanese) TK brand series that begin with an A.
Those are actually made by OST who is out of TW I think.
.
The KZG/KZJ problem is because they pushed the electrolyte too far.
They both use the same one and it's simply not stable.
BTW: Both are out of production now. Quit making them.

As for having a well cooled case, the case at issue is a steel case, not plastic. I'm not sure how much better that would conduct heat by comparison. But it has no vents (or sunction) on the sides, so I suppose that might not work out as well. As far as ventilating the board as a whole, there are only two fans in it right now - the 3500 RPM AVC (for the CPU, which won't last much longer but shall soon be replaced, yet still runs at full speed, again) and the power supply fan that's directed into the case (the graphics card also has its own fan, of course, to reiterate, though that graphics card might not last much longer either). I'm not sure how adequate that would be in this case for cooling. But I know none of the temperatures (save by the graphics card in 3D load, I estimate) ever hit 60C, much less 40C-50C. I'd hope that would mean a well ventilated case in the case of Intel 85C Nichicon caps going bad after 5 years...

Intel considers the lifespan of a system as 7 years at 8hr/day 5 days a week. That's from their marketing side not the tech side and it's largely based on how long until things [including the OS] are obsolete. Based on that the tech side builds everything to last 'at least' 7 years. Most things end up lasting a lot longer but by policy 7 years is the actual requirement. Those 85C caps would last that long with that kind of usage.

The 85c caps I speak of look like this, maybe just a might bit smaller and thinner (and might have only one VRM print instead of two):

http://media.digikey.com/photos/Nich...%20990-992.jpg

Yup, those are the ones.
I've seen those bloated on everything from socket 370 to 775.
I posted pics of bloated ones from a couple of Intel boards a long while back.
It's certainly not a 'usual' thing but it does happen enough to look out for it.

Can caps be bloated at the bottom as well (might not be an astute question)? I ask because it looks like there is more discrepancy between the bottom of the cap in width and the top in the case of my 85C caps, but it looks that way in the linked image as well (though maybe a bit less), so I'm wondering if that's just by design....

Could happen if they tried to push the bung out but I've never seen any of that particular cap fail that way.
The bottom of caps are pressed to hold the bung in and keep a good seal.

One last thing is the old power supply. I should have detailed out what happened with the ill-defunct 300-Watt PSU - if I didn't wait long enough I'd either have a flash of the LED light or the computer powering itself on and off 5-6 times in a row (within 5-6 seconds) before powering on correctly (this didn't happen too often but noticeably enough). Does that put unusual stress (or nontypical load) upon the capacitors, given that it's just before boot (by way of thermals)?

Power cycles like that shouldn't hurt caps but it might give ICs problems. You have to cycle power pretty fast and a lot more times to heat up a cap.
.
That could have happened from bad caps in the control circuits in the PSU or bad output filter caps.
The OP filter caps and mobo caps have to be charged to some level [maybe 80%] within something like 50 msec or a safety feature kicks in and shuts the PSU back off. [Specifics vary by ATX spec version.] The design assumes that if the voltage isn't coming up in that much time then there is a short somewhere, so it shuts off.

And, since I'm not too knowledgeable on this, would the ripple current be irrevocably worsened with such a power supply, possibly wreaking havoc upon the life of the capacitors (I imagine 450 watts might be a bit of an overkill for a computer like this but it was the only one I trusted that I had)?

ON-OFF cycles won't make for more ripple but failing PSU caps will.
PSUs don't 'push' power. It's 'pulled' from them by the system as needed. One with a bit of over-kill is fine and a good idea. Going nuts with over-kill is just a waste though because there won't be enough load on it to keep it in the range were it has good efficiency. With really serious over-kill it might even have regualtion stability issues.

As for problems, the closest to it is the CMOS battery failure. I'm not sure if it's a dead CMOS battery or a capacitor problem.... could it be the 85C caps going? Since, again, the CMOS error reaches occurrence the same way every time and has not caused any other issues, period.

Yes, bad small caps is one possible cause for something like that.
.

[mariushm]

Re: Question about 85c caps vs. 105c caps

As far as I know, the KZG had a problem with the ratio of chemical elements in the electrolytic substance or something like that. It was not related to the Aluminum casing being less "pure".

The tests you speak of in the first question are the ones that tell the manufacturer how many hours @ x degrees the capacitor will last.

Those tests give an estimate, basically saying if the capacitors are subjected for various periods of times at temperatures close to that 85c or 105c, when the total amount of time exceeds that value, the capacitor will probably no longer be within the specified parameters (+/- 20% capacity and so on).

The capacitor's theoretical life will increase exponentially as it's kept at lower temperatures, but 30-40c I'd say are normal temperatures for capacitors. If they're not abused by the electric circuit they're in (you can keep the pc at 0c but if a mosfet dies for some reason and sends 100v through a 6.3v rated capacitor what can you do) and there were no manufacturing problems, at such temperatures they should last decades.

About the manufacturing problems.. you can do tests and tests but sometimes it can be something as simple as poorly calibrated equipment or a batch of chemicals is not right. Sometimes, the manufacturers simply notice a higher mortality rate for a series, decide to check it out and probably issue recalls for that series.

Don't worry about the construction of the case, the fan speeds, it doesn't matter that the CPU or video card's chips have certain temperatures. What matters for capacitors is the temperature of the air surrounding them and the temperature of the PCB board they're soldered to. As you have fans in the system, the air gets moved around so the capacitors are not exposed to the 50-60c the CPU is exposed to.

Yeah, caps can be bloated only at the bottom. Small caps generally don't have those "cuts" at the top because there's little danger of them exploding, yet sometimes gas gets produced inside and as the bottom is the only release...

[Wester547]

Re: Question about 85c caps vs. 105c caps

PCBONEZ, thank you for the response. I find it very informative. I apologize for the length again - I just wanted to be as informative as I could be.

Points of discussion I'd like to touch base on:

Is there an industry standard test performed on all caps/boards before they're put into production (and are released in motherboards) that affects the lifespan of all boards inevitably from day one (if I read correctly)?

Chinese aluminum, are you sure no Japanese brands have Taiwan aluminum foil? I'm wondering about that in respect to KZG caps - I recollect that they're from Chemicon (Japan) yet they still die prematurely (after maybe 4-5 years of use). They don't have Taiwan aluminum?

As for having a well cooled case, the case at issue is a steel case, not plastic. I'm not sure how much better that would conduct heat by comparison. But it has no vents (or sunction) on the sides, so I suppose that might not work out as well. As far as ventilating the board as a whole, there are only two fans in it right now - the 3500 RPM AVC (for the CPU, which won't last much longer but shall soon be replaced, yet still runs at full speed, again) and the power supply fan that's directed into the case (the graphics card also has its own fan, of course, to reiterate, though that graphics card might not last much longer either). I'm not sure how adequate that would be in this case for cooling. But I know none of the temperatures (save by the graphics card in 3D load, I estimate) ever hit 60C, much less 40C-50C. I'd hope that would mean a well ventilated case in the case of Intel 85C Nichicon caps going bad after 5 years...

The 85c caps I speak of look like this, maybe just a might bit smaller and thinner (and might have only one VRM print instead of two):

http://media.digikey.com/photos/Nich...%20990-992.jpg

But they're more thin than small, I think, and even not terribly so. Can caps be bloated at only the bottom as well (might not be an astute question)? I ask because it looks like there is more discrepancy between the bottom of the cap in width and the top in the case of my 85C caps, but it looks that way in the linked image as well (though maybe a bit less), so I'm wondering if that's just by design.... (the PS polymers might also be just solid aluminum, my mistake ^^; )

One last thing is the old power supply. I should have detailed out what happened with the ill-defunct 300-Watt PSU - if I didn't wait long enough I'd either have a flash of the LED light (and failure to boot any) or the computer powering itself on and off 5-6 times in a row (within 5-6 seconds) before powering on correctly (this didn't happen too often but noticeably enough). Does that put unusual stress (or nontypical load) upon the capacitors, given that it's just before boot (by way of thermals)? And, since I'm not too knowledgeable on this, would the ripple current be irrevocably worsened with such a power supply, possibly wreaking havoc upon the life of the capacitors (I imagine the 450 watt replacement might be a bit of an overkill for a computer like this but it was the only one I trusted that I had)?

As for problems, the closest thing is the CMOS battery failure. I'm not sure if it's a dead CMOS battery or a capacitor problem.... could it be the 85C caps going? Since, again, the CMOS error (checksum) reaches occurrence the same way every time and has not caused any other issues, period.

[PCBONEZ]

Re: Question about 85c caps vs. 105c caps

If these 85C are the little black VR then they do bloat but because they are so small it is very hard to see it.
-
If you aren't having issues and they aren't bloated I wouldn't worry about them.
-
If you are having weird problems then change them.
.

[PCBONEZ]

Re: Question about 85c caps vs. 105c caps

There is no 5 year rule I know of.

The wikipedia article on caps plague has been screwed up since day one.

The hours in a data sheet is NOT estimated Lifespan.
The correct name is Endurance Lifetime.
[Short version] It's how long it can have the snot abused out of it [specific conditions in an industry standard test] and still work fine.
NOTHING says how long it is expected to last after the test......
Actual Lifespan is generally like 10-20 years.
~~~~ However:
All the factors that affect the length of the Endurance Test also affect actual Lifespan by the same percent.
The biggie is reducing temp by 10C will double both the cap's Load Life -and- it's Lifespan.

[Considering the last sentence]
A 85C @2000Hr cap [theoretically] will last 25% as long as a 105C @2000Hr cap.
105C @2000Hr cap [With doubling for each 10C drop]
95C - 4000Hr
85C - 8000Hr - Compare to the 85C @2000Hr cap
[I said 'theoretically' because this only works if you assume the 105C and 85C have the same estimated LifeSPAN and they probably don't.]

No manufacturer publishes an estimated LifeSPAN.
That could get them into legal problems so they just don't.

The P50/P51 problem
- Has been gone since 2004 or 2005.
- Affected NO Japanese brands.
- Did not have anything to do with the run of bad Nichicon HM and HN. (Which are Japanese.)
- Does not affect -any- Intel brand board because they only used Japanese caps back then.

Caps that use Chinese aluminum [almost all CH and TW made caps] have other issues that cause the same problem as the P50/P51 formula, it's just slower.
[Specifically CH mined aluminum has different and more impurities and they cause chemical changes in the electrolytes.]
- THIS problem has NOT gone away.
- No Japanese brand uses CH mined aluminum.

Intel does occasionally use 85C [usually Nichicon] caps that will start going bad after 4 or 5 years old.
It depends a great deal on how many hours the system has been running and on how well the system is cooled.
A board in a well cooled case may never have that problem at all.
.
.
If I missed anything critical ask again.
- Cause you wrote A LOT for one post and I only skimmed through it...
.

[Wester547]

Re: Question about 85c caps vs. 105c caps

Sorry for the lack of information - I live in California, if that helps. It was never on for 24/7, as stated. I also apologize for the elongated post - I just want to give as much information as I can, should anyone have an answer. ^^;

[selldoor]

Re: Question about 85c caps vs. 105c caps

Its too long.

Nobody knows.

I'd guess few care.

read this: its even longer.http://en.wikipedia.org/wiki/Electrolytic_capacitor

Having skimmed you dont even tell us where you live.
Sahara? Alaska hot cold dry damp- You have no idea of the history of the machine. If it was on for a year off for two and on 24/7 for one would produce a totally different result than if on 10/7 fo 4 years.

Like I said TOOOOOOOOO Long and NOBODY knows.
Just save up for a new one.

[Wester547]

Hello everyone. I registered here a bit ago but I have a few questions regarding capacitor lifetime and rated temperatures with regard to that. I know this has been explained on these boards before, so if I'm asking banal questions or asking something that's already been answered, please don't hesitate to tell me. ^^; I also hope this is posted to the right subforum.

Recently, in my family, I was given an old computer to have and mess with. It's been around since January of 2003 but I suspect that it has no more than 4 years of usage on it (maybe 3 at the least, not 24/7 but closer to half that). It's a VPR Matrix FT7100-PE that has a D845PEBT2 motherboard by Intel, a Pentium 4 2.66 GHz, 1GB of single channel, PC333 DDR-SDRAM, a 128MB DDR Radeon 9500 Pro, a Sound Blaster Audigy gamer, two Pioneer 16x DVD-RW drives, and a ST3250824A and ST3160023A (with the ST3250824A being the OS drive), along with a 450-watt BFG power supply and Windows XP Professional Edition SP3 installed. The power supply in question recently replaced a garbage (FDD) 300-Watt PSU that started forebode its own death a year earlier (but didn't quite die yet) when it took waiting a couple minutes after turning on the power strip for the system to power up properly. I say I am of the belief that the system has had 4 years of use because the WD800JB in there that recently died had around 360 days of use on it judging by the S.M.A.R.T. data report (and this was closer to being a 10/7 machine than a 24/7 one, as aforementioned, though it's only had about 2 years of use since February of 2006 since that was when the ST3160023A first had its own installation of Windows XP, the only OS that was ever on the computer).

The ST3160023A only has 237 days of use (and ~3800 power cycles), and the WD800JB was bought in early 2005 to replace the defunct 120GB PATA Western Digital drive that came with the machine. So I would conjecture that 4 years of use is about right (given that the computer was used lots from 2003-2005). I recently installed the power supply and ST3250824A (which has almost 20,000 hours/~2900 power cycles from it and is from a now dead computer [the 450-watt BFG PSU is as well, and has a couple years of use on it], but still works well, in spite of having a few recently unearthed bad sectors, 8 reallocated and one offline uncorrectable/pending, as well as 9 reported/uncorrected errors) As obvious, it's a rather archaic machine, but I imagine it still suffices for light tasks, such as word processing and web browsing (Microsoft Word 2003/Open Office and Google Chrome respectively). ^^; And perhaps occasional old games. But all this verbose aside, what I mean to ask is how long the knowledgeable people here think this machine will survive, because whilst, to my relief, most of the caps on the board are 105C Nichicon HD caps (1000uf @ 16V seems to be the Nichicon specs, and there are also 2200uf @ 25v caps on the board, along with some 100uf caps and 470uf ones, from what I can see) or United Chemicon PS caps (105C, and they're solid electrolytes rather than liquid, I think), there are a couple 85C caps on the board, both of which are probably Nichicon VRMs... 105C caps would clearly quadruple the lifetime of 85C caps.

I am relieved that it seems that this motherboard's capacitors are not beleaguered with either the P50/P51 electrolyte plague, nor the defect with the Nichicon HM/HN capacitors of the earlier 2000's. And I don't think they'd have Taiwan Aluminum foil either, as to my knowledge, Nichicon processes most of their aluminum foil in-house and generally do not have plants in Taiwan, and all capacitors from Chemicon, aside from KZGs or caps that aren't made to last long, are solid (I believe; KZGs killed the machine the ST3250824A was in initially, though that machine wasn't to last with a nForce 6100 chipset). The only thing holding it back is the couple 85C caps. I have read on this forum that 85C caps don't last much longer than 5 years... this is 5 years of 10/7 use, right? Because this computer has been power cycled around 6000-7000 times, I would imagine (knowing the WD800JB had almost 6,000 power cycles). Would power cycling (or thermal expansion) be as harmful to capacitors as lots of power on hours and sustained heat, or less? Or more, or just as much? I would imagine that power cycling affects all electronics, but I'm just curious to get a more professional viewpoint.

As for the system's temperatures, the Pentium 4 has a 3500RPM fan and the BFG PSU has a fan directed inward to the system (and, of course, the 9500 Pro has its own fan), where as the original PSU did not. Haplessly, the 3500RPM fan's ball bearings don't sound good, even though it still runs at full speed (nor is the fan speed regulated so it's always at 100%) - it's been that way for a year. No surprise, I guess - it's an AVC fan and they aren't well reputed from what I read. But I suppose I could find a suitable replacement for it (given that it's Socket 478) once the fan goes. So, I would think that the AVC fan aside, and given future excursions, that the system is ventilated adequately enough. But the temperatures themselves, for the CPU (and the core itself), are generally a bit under 40C at idle and around 55C at load (on a fairly warm day, but that's with 3D acceleration on the 9500 Pro - without, it's closer to 50C), and the Auxiliary and mainboard are in the 30's. The 9500 Pro has no temperature sensor but I don't remember reading about it being a hot card, and in its lifespan it's mostly been used for 2D work (though I guess the fan on it could stop working at any moment). The ST3250824A is usually around 37C on such a day and the ST3160023A 39C-40C (but it was once as hot as 58C during a short offline test with regard to S.M.A.R.T.... also, none of the WD drives had a temperature sensor, but the WD800JB was right beneath the ST3160023A, as was the 120GB), though they are not loaded heavily so their temperatures are rather solidified, and the drives are not atop one another, either. I should also make of note that the enclosure for the HDDs seems to be the farthest away from the rest of the components.

Does anyone here think given the 85C capacitors on the board that the 5 year ruling applies to this system or could it last a bit longer? The only other thing wrong with the system seems to be a dead CMOS battery for the second time. I'm wondering whether that's a ramification of a dead battery or dying capacitors (though there are utterly no indications of it otherwise and the BIOS shows the same exact CMOS error every time). PCBONEZ posted a concise and good reference for how long 105C caps by comparison to 85C caps. But was this temperature guide alluding to load temperatures or idle temperatures? Or average temperatures, with relation to the heat the rest of the components generate? ^^; Also, do 85C electrolytic caps dry out faster 105c electrolytic caps (by a significant amount) regardless of use (since I've garnered that electrolytic caps inevitably dry out after 15-20 years, similar to batteries)? There are also no visual indications that any of the capacitors on the motherboard are going, nor do I smell or hear such a thing ever.

I hope this wasn't too long a post to read nor too much of a muss. It's not life or death if this computer doesn't last much longer (nor do I think I would be given allowance to recap it or have it recapped), I just always find this kind of knowledge fascinating. I appreciate any feedback muchly. (: