Rubycon MBZ leaks in storage!

[goodpsusearch]

I removed some good caps from a Dell Optiplex GX260 mb 4 years ago.

Past from visual inspection, I also tested them with esr micro and their ESR and uF were good.

Yesterday I noticed one of them was leaking and bulging.

Its current esr is 0.85 ohm and the capacity 824uF.

Those caps have been stored in a dry place and the temperature varied from 15°C to no more than 25°C

The other cap is still good. I don't know for how much though..

[Uranium-235]

Re: Rubycon MBZ leaks in storage!

even in storage a used cap is a used cap, sometimes this happens

[Wester547]

Re: Rubycon MBZ leaks in storage!

....

Well, that's scary...

Even scarier, it's an 8x16 1000uF 10V MBZ, the same kind of MBZ Topcat pulled from those Thinkcentres three years ago, and most of them failed without any signs. I chalked that up to a bad run of bungs at the time (or someone who abused the leads whilst installing them), but now I'm not so sure... from what I can see, the capacitor was manufactured during the 13th week of 2002, in Singapore (March 2002... I think "A" means Singapore because I've seen 1000uF 6.3v 8x11.5 YXGs in a sealed bag before with "A" codes, and the label on the bag said that it came from Rubycon's Singapore factory). They are well over 13 years old at this point, though, so it might just be their time (even non-aqueous lytics have about a 20-30 year shelf life). Sometimes defects in the electrolyte take many years to cause issues and MBZ is only a 2,000 hour @ 105*C rated capacitor.

Did it come out of a SFF GX260? Perhaps it was exposed to an excess of heat for too long and hence the bulging is the long-term effect (like what happens when you reflow Xbox 360 motherboards with the lytics still on). Also note that no capacitor is impeccable. When you produce millions of them, some are bound to be bad even from the factory. Another explanation is that once you get to MBZ-grade capacitors that lytics are no longer completely reliable. Think about how unreliable HMs once were and how unreliable KZGs are. Sometimes even FJs and WGs fail for no reason.... there is a reason why lytics don't go lower than HZ grade. It is very difficult to increase the conductivity of the electrolyte by way of ions. The increase in the H2O in the electrolyte necessary to improve conductivity makes all capacitors more sensitive to heat and that kind of electrolyte has a tendency to aggressively assault the dielectric until it's gone and hydrogen gas forms, which the proper neutralizers will reduce the likelihood of (which is why polymers exist - conduction by way of electrons, hence the solid state terminology). And that kind of electrolyte in a high concentration dries up faster than lytics with "conventional" electrolyte as well.

There is also a possibility that the electrolyte completely broke down after the capacitors just sat there for four years. The aluminum oxide layer thins as long as the capacitor sits in storage with no voltage across the plates and no current going through it. That's not good for any capacitor. I think Panasonic recommends that 1 year be the maximum time lytics sit in storage without reforming. But maybe the simplest solution to come to is, don't trust those 8x16 1000uF 10V MBZs.

[stj]

Re: Rubycon MBZ leaks in storage!

i'v seen pana FC's leak around the leads,

you have to remember that rubber goes hard with age.
*any* cap with a rubber bung will have it's seal comromised eventually.

interestingly i'v had old soviet caps that probably dont have this problem,
the bung is plastic and is inserted with a rubber seal at the crimp-point - so it cant leak around the pins.

i'v also seen old american caps from the late 70's and early 80's where the bung has an epoxy cap over it to prevent leaks. ( "T.I." branded)

[TechGeek]

Re: Rubycon MBZ leaks in storage!

i'v seen pana FC's leak around the leads,

you have to remember that rubber goes hard with age.
*any* cap with a rubber bung will have it's seal comromised eventually.

interestingly i'v had old soviet caps that probably dont have this problem,
the bung is plastic and is inserted with a rubber seal at the crimp-point - so it cant leak around the pins.

i'v also seen old american caps from the late 70's and early 80's where the bung has an epoxy cap over it to prevent leaks. ( "T.I." branded)

Those were the good days. Made in America, high quality, not the Chinese junk we have today.

[Pentium4]

Re: Rubycon MBZ leaks in storage!

I'd give the thing a break. It's 13 years old now. It probably got quite a bit of use, and then 4 years is a long time to sit for an already almost decade old capacitor. This is why I'm always a little skeptical about reusing caps.

[mockingbird]

Re: Rubycon MBZ leaks in storage!

I've got some Rubycon MBZ/MCZ, specifically:

MBZ
6.3v 2200uF:
0340
0227
0202
0202
16v 1800uF:
0333
0333
0333
16v 1500uF:
0227

MCZ
6.3v 1500uF
0542
0542
0542
6.3v 820uF
0521
0521

None of them look defective, nor do I have any reason to believe they are.

All the datecodes are prefixed with "T".

I think Wester547 is on to something when he mentioned the Singapore factory (datecode prefixed with "A").

[Wester547]

Re: Rubycon MBZ leaks in storage!

^ I don't think that's it because if you look at Topcat's thread of failed MBZs, the 8x16 1000uF 10V ones, that failed without showing it, had "T" prefixes (Japan). I agree with P4. There is a reason why manufactures recommend against reusing old capacitors. And I have a hunch that as the dielectric thins on the shelf, the neutralizers lose their effectualness in protecting the dielectric from the aggressiveness of the hugely water-based electrolyte (they can't really protect something that just thins and thins). That could as well be what happened.

And IMO, the factory from which they came really isn't a way to gauge at the quality of certain series. I've seen KZGs and KZJs fail prematurely from almost every factory I've ever seen labeled on Chemi-con sleeves, for an example... "(k)" - Indonesia. "(T)" - Japan. "(2)" - probably Japan. "(3)", "(7)", etc...

Also agree with P4 with respect to the age of the capacitors. 2,000 hours @ 105*C capacitors just don't last long as series with superior endurance and I don't think they're meant to (even though some are obviously better than others).

[mockingbird]

Re: Rubycon MBZ leaks in storage!

Would you say reforming these types of caps would prevent failure then, no matter the age?

Also, consider that in P4 machines which often had terrible PSUs and poor ventilation, MBZ rarely or never bloated in them, even while KZG/KZJ spewed their guts. This 2006'ish Foxconn board is a good example (MBZ on VRM high, KZG on VRM low).

With regards to Topcat's thread, For some reason I've always had a suspicion in the back of my head that Rubycon either intentionally or otherwise produced some inferior batches, which went out to Lenovo and Dell.

Hopefully I'll have the means soon to measure these caps (With a host of a whole bunch of others) to confirm that they're still good, with a product I expect to receive from our great forum member Behemot.

[stj]

Re: Rubycon MBZ leaks in storage!

it could just be down to how they ran before being salvaged.

[Wester547]

Re: Rubycon MBZ leaks in storage!

Would you say reforming these types of caps would prevent failure then, no matter the age?

I couldn't tell you. I don't think goodpsusearch's findings reflect well on Rubycon MBZ, because bloating some time after being pulled is the kind of thing KZGs and KZJs do.... and bad HMs and HNs. And Fuhjyyus and CapXons. On the other hand, those MBZs could have had a rough life and are over 13 years old, so I wouldn't rail Rubycon on for it. Most of the time, MBZ is a very reliable series and handles heat quite a bit better than MCZ.

Also, consider that in P4 machines which often had terrible PSUs and poor ventilation, MBZ rarely or never bloated in them, even while KZG/KZJ spewed their guts. This 2006'ish Foxconn board is a good example (MBZ on VRM high, KZG on VRM low).

Looks to me like it has KZJ in the VRM output, but I doubt KZG would do any better - maybe even worse. Those 3300uF 6.3V KZGs are really the worst of them all. Interesting thing about VRM in/out failures.... most of the time the failed KZGs and KZJs, or the ones that failed the fastest, are on the lower voltage rails and not the higher voltages one unless they run too hot for too long, even though KZGs do have a tendency to silently fail. That's because (I think) the higher voltages rebuild the oxide layer faster and prolongs failure (though it consumes more electrolyte in the process since those KZGs/Js are working closer to their maximum voltage rating).

With regards to Topcat's thread, For some reason I've always had a suspicion in the back of my head that Rubycon either intentionally or otherwise produced some inferior batches, which went out to Lenovo and Dell.

Makes me wonder if KZG and KZJ are intentionally bad! But I honestly think those "good" companies have better standards than that. In the case of those MBZs, since most of them failed without showing it, it was either A) a bad run of bungs or B) someone didn't treat the capacitor leads with much care, compromising the quality of the seals. Or.... C) 1000uf 10V Rubycon MBZs of the 8x16 dimensions might have some chemistry issues are are best avoided.

[mockingbird]

Re: Rubycon MBZ leaks in storage!

Looks to me like it has KZJ in the VRM output, but I doubt KZG would do any better - maybe even worse.

Oops. Yup, that's my board... Outlasted the re-capped fleamarket SP-450 PSU, decided to repurpose it as a server in a school, after I upgraded to my poor man's quad-core.

I must say fascinating theory on the behaviour of failure on KZG/KZJ... Kudos for that.

Makes me wonder if KZG and KZJ are intentionally bad!

The better question is, did UCC ever actually find out a suitable electrolyte formula for ultra-low ESR caps. Someone put forward recently in a thread here that it's even possible until the very end of the series' run that UCC never even knew about the premature failures. Considering Japanese culture, I can imagine a company like UCC being very stubborn about the quality of their product, given their track record. Foolish of them though if you ask me, because they've had several dud series before.

Then again, it could have been intentional, as you said. Planned obsolescence and all that.

But I honestly think those "good" companies have better standards than that.

To be fair to them, UCC is the only Japanese company I can think of that didn't outright buy another polymer cap manufacturer like Panasonic buying Sanyo's polymer division or Nichicon buying Fujicap. You gotta give props to them for that.

[Wester547]

Re: Rubycon MBZ leaks in storage!

I must say fascinating theory on the behaviour of failure on KZG/KZJ... Kudos for that.

I don't know if that's really correct, but I think it's just simple logic. I have noticed that behavior with many bad cap series - the bad HMs, and even Teapo, etc.

The better question is, did UCC ever actually find out a suitable electrolyte formula for ultra-low ESR caps.

The baffling thing about that is that if you look at UCC's KZG datasheet (the one by UCC and not NCC), it is said that KZE use the same electrolyte technology as KZG and KZJ! And KY's datasheet further says that KZE and KY use the same electrolyte too. So it must not be inherently bad electrolyte, but using it (their foray into water-base electrolyte) in greater concentration must cause the electrolyte to break down and attack the aluminum foil, which makes sense to me. KZE and KY have an extremely good track record compared to KZG and KZJ, so I don't know what else to make of it.

Someone put forward recently in a thread here that it's even possible until the very end of the series' run that UCC never even knew about the premature failures. Considering Japanese culture, I can imagine a company like UCC being very stubborn about the quality of their product, given their track record. Foolish of them though if you ask me, because they've had several dud series before

I don't consider those old series bad at all. The fact that they almost never bulge and that the bungs almost always give away and allow for electrolyte to leak quite simply means that the bungs hardened and decomposed over time, like stj said. Those capacitors do not have bad electrolyte by a long shot. The fact that the bungs decomposed first goes to show you how good the electrolyte is, and I think Rubycon, Nichicon, Chemicon, and Panasonic have improved upon the quality of their seals over time, but as stj said, the rubber bungs do eventually decompose and end the life of the capacitor.

To be fair to them, UCC is the only Japanese company I can think of that didn't outright buy another polymer cap manufacturer like Panasonic buying Sanyo's polymer division or Nichicon buying Fujicap. You gotta give props to them for that.

No, but UCC bought Marcon's divison of lytics many years ago, 1995 IIRC.

I think Rubycon had an advantage over UCC and some Nichicon in terms of ultra low ESR lytics because of something theOracle said some time ago. He said that a friend, in the know as an engineer at Sony, had noticed that Rubycon's primary plant in Japan borders on a spring of incredible purity. Rubycon must have a special kind of water-base electrolyte that yields much better conductivity so they can make ultra low ESR lytics with less of it, which means they don't fail as much and have better stability characteristics. Just a theory, I don't know if it's really true. Similar may be true of Panasonic - their capacitors are expensive for a reason so they must use very expensive materials, electrolyte included, and that electrolyte probably has better conductivity than their peers, giving them another edge over the competition. Again, just a theory.

This thread though is proof that even Rubycon MBZ isn't invincible, and in that sense all electrolytics will eventually fail, even the good ones.

[goodpsusearch]

Re: Rubycon MBZ leaks in storage!

I reuse caps all the time with no issues till now.

On the other hand I have seen several caps fail in storage, mainly KZG, KZJ, and bad caps. It's like capacitors prefer getting some use every now and then (ripple, DC voltage and heat) rather than staying for months unused.

[ben7]

Re: Rubycon MBZ leaks in storage!

I think I have a few 1800uF 16V (10mm dia) rubycon MBZ with the A on the datecode and they are from 2002 if I recall correctly. Will have to check later today.

I couldn't tell you. I don't think goodpsusearch's findings reflect well on Rubycon MBZ, because bloating some time after being pulled is the kind of thing KZGs and KZJs do.... and bad HMs and HNs. And Fuhjyyus and CapXons. On the other hand, those MBZs could have had a rough life and are over 13 years old, so I wouldn't rail Rubycon on for it. Most of the time, MBZ is a very reliable series and handles heat quite a bit better than MCZ.

Looks to me like it has KZJ in the VRM output, but I doubt KZG would do any better - maybe even worse. Those 3300uF 6.3V KZGs are really the worst of them all. Interesting thing about VRM in/out failures.... most of the time the failed KZGs and KZJs, or the ones that failed the fastest, are on the lower voltage rails and not the higher voltages one unless they run too hot for too long, even though KZGs do have a tendency to silently fail. That's because (I think) the higher voltages rebuild the oxide layer faster and prolongs failure (though it consumes more electrolyte in the process since those KZGs/Js are working closer to their maximum voltage rating).

Makes me wonder if KZG and KZJ are intentionally bad! But I honestly think those "good" companies have better standards than that. In the case of those MBZs, since most of them failed without showing it, it was either A) a bad run of bungs or B) someone didn't treat the capacitor leads with much care, compromising the quality of the seals. Or.... C) 1000uf 10V Rubycon MBZs of the 8x16 dimensions might have some chemistry issues are are best avoided.

Uh, keep in mind, they (KZG) are only rated for 2000 hours. Even if the caps last less than a year, as long as they have lasted over 2000 hours, then I don't really see any issue. Obviously, the 2000 hour rating doesn't mean that they won't work at 2001 hours ... it's nice to have them last longer than 2000 hours, which some capacitor series will.

[stj]

Re: Rubycon MBZ leaks in storage!

that's 2000 hours at 105'
the life increases as you decrease the temp.

[Wester547]

Re: Rubycon MBZ leaks in storage!

Uh, keep in mind, they (KZG) are only rated for 2000 hours. Even if the caps last less than a year, as long as they have lasted over 2000 hours, then I don't really see any issue. Obviously, the 2000 hour rating doesn't mean that they won't work at 2001 hours ... it's nice to have them last longer than 2000 hours, which some capacitor series will.

The very first version of NCC KZG's datasheet that I posted in the "What pisses you off?" thread rates the 10mm 16V KZGs (10x12.5 680uF 16V, 10x16 1000uF 16V, 10x20 1500uF 16V, 10x25 1800uF 16V) for 3,000 hours.

---

Well, at 65*C, they are supposed to last 32,000 hours and at 60*C, 45,200 hours according to the "10*C decrease doubles the life" rule. Most of the time, they don't last anywhere near that long at lower temperatures. I think bad capacitors are too unpredictable to follow that particular ruling. Furthermore, I've seen 2,000 hour @ 85*C and 1,000 hour @ 105*C GP capacitors outlast 2,000 hour @ 105*C KZGs and KZJs, even in storage. So something is definitely wrong with them. I don't really think it's about the endurance rating, though, it's about the individual solution each capacitor uses. And this thread makes me think twice about knocking them (KZGs) - I guess all ultra low ESR capacitors have the same issue, but that for some it just takes longer for those issues to surface, even if it's many years...

We all know that if capacitors hate anything the most, it's heat. Heat is the ultimate enemy of all lytics and greatly expedites any chemistry issues any capacitor may have. I think they'd be more likely to fail if you heated them right up than if you let them sit. That said, it's probably a better bet to use them and keep them cool above all else.

[mockingbird]

Re: Rubycon MBZ leaks in storage!

Tne very first version of NCC KZG's datasheet that I posted in the "What pisses you off?" thread rates the 10mm 16V KZGs (10x12.5 680uF 16V, 10x16 1000uF 16V, 10x20 1500uF 16V, 10x25 1800uF 16V) for 3,000 hours.

Interesting. I did a double-take when I read this, and looked my own KZG datasheet. I found the one you were refering to in this post. So now I have three of them:

I guess all ultra low ESR capacitors have the same issue, but that for some it just takes longer for those issues to surface, even if it's many years...

What do you think about Panasonic ultra low ESR? I've never seen reports of defective Samxon GC/GD/GA either (Though I think they bought their electrolyte from one of the Japanese vendors). What about Sanyo WG?

[stj]

Re: Rubycon MBZ leaks in storage!

pana dont make ultra-low esr caps, FR is the lowest they have.

(i'm not including FL because you cant buy them, and there is no official datasheet)

[Wester547]

Re: Rubycon MBZ leaks in storage!

Panasonic do make ultra low ESR capacitors, they just sell them directly to OEMs. FJ (and FJS/FJZ I think) are equal to HM, the datasheet for FJ does exist (you can find it if you Google it). Don't know about FL but in accordance to an old post by PCBONEZ, they should be equal to FJ, according to an old ad he saw by Panasonic.

What do you think about Panasonic ultra low ESR? I've never seen reports of defective Samxon GC/GD/GA either (Though I think they bought their electrolyte from one of the Japanese vendors). What about Sanyo WG?

Actually, premature failures of Samxon GC/GD definitely exist...

https://www.badcaps.net/forum/showpo...&postcount=164
https://www.badcaps.net/forum/showpo...29&postcount=3
https://www.badcaps.net/forum/showpo...9&postcount=42
https://www.badcaps.net/forum/showpo...&postcount=167
https://www.badcaps.net/forum/showpo...&postcount=172

And a few posts in this thread:

https://www.badcaps.net/forum/showth...=gc#post414522

Haven't seen a GA fail prematurely yet, but it's hard to say. And we all know how unreliable Samxon GF and GK are, and they are equal to HE and HD respectively, so I don't think that can bode well for them either. Panasonic ultra low ESR capacitors are great, in my experience - no complaints. Sanyo WG are also good, though I've found them more resistant to heat than the WX series (equal to HD) which always bewildered me. WF (equal to HN), not so good, but Sanyo capacitors, at least on Intel/Foxconn/Dell motherboards, are seldom seen compared to the other brands so it's hard for me to form a conclusive opinion about them, so I won't call them out because they haven't caused the premature death of millions of motherboards like KZG have.