Bad Capacitors A7N8X Deluxe

[Gibby]

I edited an earlier reply to remove a statement about the A7N8X using the 12v rail for CPU power. ASUS nForce2 motherboards (and one other manufacturer - MSI if I recall) used the 5V rail for CPU power, while most other motherboard manufacturers' nForce2 / Socket A boards had switched to the 12V rail.

They didn't use the power directly, but converted it onboard through the CPU's voltage regulatoion module (VRM). The VRM is just a DC-DC converter, and theoretically could be designed to use almost any input voltage.

There is no absolute better or worse for using 5V or 12V. Either one had it's downside related to the bad cap problem, but for different reasons.

In the case of the A7N8X, the two-phase VRM module and it's power capacitors were running pretty close to the design limits after people started using the newer, power-hungry Athlons (back in 2003). Heat from the CPU and northbridge stressed the components even more. Toss in some capacitors with bad electrolyte and it was a disaster waiting to happen.

But there were also bad capacitors that found their way into many PSUs. Most notable were the Antec 430 and 480 watt units using Fuhjyyu (yes, that's the correct spelling...) capacitors. When those capacitors started going, people would most often see unusually low voltage on the 5V rail. If your BIOS reports less than 4.6V on the 5V rail right after a cold boot, then you probably have this problem. It was not limited to either Antec PSUs or Fuhjyyu caps - those were just the most frequently seen in forums.

A bad cap in your PSU looks the same as the ones shown on TC's motherboard. Leaking electrolyte from the top or bottom of the cap, or a puffed-out looking case on the side or top. Caution: ripping open your PSU to examine the caps will void the warrantee. If it's an Antec TruePower 430, it's either dead already or out of warantee. I don't think Antec ever admitted the problem or offered to replace the defective units (except with another TP430, but even then you got stuck with shipping).

The VRM on any motherboard tries to keep it's ouput - the CPU voltage - steady no matter what the input voltage actually is. It will still try to do this even if the 5V rail is only 4.5V, but it does this at the expense of increased VRM heat and loss of stability. So pair an A7N8X with a TruePower 430 (which was pretty common a few years back thanks to all the Antec fanboy reviews) and you can understand how this would be an issue.

Now take this whole setup and toss it in one of those nice new (~2004) Antec Sonata cases - one of the early ones without all the extra vent holes drilled into the side of the case. Yup, the ones with the poor airflow that cooked the motherboard. Sprinkle in a touch of overclocking and you have pretty much described my home computer (at the time). Take out the overclocking and you have pretty much described 250+ machines we had at work. This doesn't make me any kind of expert, just someone who is painfully familiar with the issue.

For the non-ASUS nForce2 motherboards, the bad cap issue was made worse by using the 12V rail for CPU power. This was mostly frequently seen in gaming rigs with higher-end Athlons, the newer (again, ~2004) powerful video cards, lots of spinny things running in the case, and early ATX 1.x PSUs with underpowered 12V rails. A stressed 12V rail putting out less than 11.5 or so volts would tax the VRM, making it more likely to cook off a marginal cap. Put that in a case with poor cooling or us a PSU with bad caps and you end up with the same kind of reliability issues that some of the A7N8X people had.

I'm not trying to slam ASUS, Nichicon, Antec or Fuhjyyu here. This is just to point out that there is a whole rat's nest of issues that were directly or indirectly related to the bad cap problem. In TC's case, there's no question that the caps are failing. But answering questions like "Why is my computer locking up / randomly rebooting" or "How come I can't overclock using XXX settings" cannot just be answered with a simple "Because you're using Brand X caps and everyone says they're bad..." I

'm sure there are plenty of A7N8Xs with Nichicon caps running in a Sonata case with a TruePower 430W PSU that have never had a problem to this day.

Well, maybe not plenty, but ther has to be some still running out there. I mean, what are the chances...

[Tech_Cowboy]

...I would look at your power suppy, your getting dirty power, thats what the caps filter out is irregular power crap and noise.
Also capacitors are good to 105c, so its not the heat, its the dirty power, probably from overclocking, spurious interference.

I've heard that the Antec 550w that I'm using is a decent PSU. But I'm sure you're right that overclocking stresses the capacitors (among other components). It's good to know that the caps are rated for such high temperatures.

TC

[Tech_Cowboy]

If you vacuum your cpu be real careful not to rock it or bump it too hard, especially if you have a huge heatsink, you'll crack the cpu real easy.

Oh yes. I'm very careful. I've done it enough that I know what I'm doing. The first thing I do is remove the fan from the heatsink and vacuum the heat sink. Most of the dust seems to get trapped right there. The reason I have to vacuum so frequently is because my house is a fairly dusty environment. That's one of the downsides of living on a ranch in Texas. No matter what you do, the dirt just seems to get everywhere.

TC

[Tech_Cowboy]

In TC's case, there's no question that the caps are failing. But answering questions like "Why is my computer locking up / randomly rebooting" or "How come I can't overclock using XXX settings" cannot just be answered with a simple "Because you're using Brand X caps and everyone says they're bad..."

Gibby,

You seem to know so much about this motherboard and electronics in general, so I hope you can help me one more time. I'm not having much luck finding those 6.3v Rubycons. Do you know where I can get them, either online or in a retail store? If I can't find them, would you recommend some other brand/manufacturer?

Thanks,
TC

[Gibby]

TC - Don't get too hung up on the Rubycon brand - we're talking about a problem with caps manufactured in the middle of 2002. In hindsight, the slightly more expensive Rubycons ended up being worth the money back then. Nichicon and all the other cap makers that had problems lost millions. They didn't just keep making their caps the same way with the same electrolyte.

Now, while I wouldn't go out of my way to hunt down Nichicons for recapping, I would probably still use them if that's all I had or all I could find. How could I be sure they were not manufactured in the summer of 2002? I just wouldn't worry about it, because they've been cranking out millions of caps since then and it's pretty unlikely that I'm going to even get one that old. How would I know that 'trusted' Japanese manufacturer's caps (vs. no-name Chinese) ones made today are actually good? I don't - I can't see into the future. I simply hope they're good because they usually have been. But Nichicon's had been good caps until the bad batch surfaced in 2003, and they're supposedly good today, so who knows.

As far as replacements, I would trust whatever brands/lines the guy at badcaps is selling. I think he has Samaxons and Rubycons. He will know the right replacement if you give him the motherboard model, and he's not going to sell you the wrong kind of cap, i.e., a type that shouldn't be used on a motherboard.

I'm not sure how you feel about eBay, but if you search their auctions for "3300uf ESR" you should get a few listings back. In general, you're looking for a manufacturer's very/super/untra-low ESR line. They will usually say right in the listing if that manufacturer's particular line suitable for motherboards. Now, can you believe them? Well, it's eBay...

I got back a listing with Panasonic, Sanyo, OST and Nichicons. Panasonic and Sanyos are probably more 'trusted', but like I said - I would still use a Nichicon today if that's all I could find.

You need 6.3v rating. You could go with a higher value, but it's unlikely that you could find a 10 or 16v one in as small of a can as you need. Can = package = physical form of the capacitor.

Can size: you need to find one in a 10x20 or 10x23 size. That's 10mm diameter by 20 or 23 mm high - probably what you have now. 12mm diameter is the other common 'size' but you won't be able to mount those flush against the motherboard like your current caps. You would have to bend the leads a little on a 12mm can to line up with the holes on your motherboard and they would sit a few mm higher than they should. I also don't think you could phsically squeeze them next to each other like you need to on this motherboard. A 10mm can's leads will be spaced perfectly for your motherboard. The height doesn't matter - you just want to avoid the 'short' 10x10 size. They're very expensive and usually have a lower service life.

You need a radial lead cap - it has both leads on one end of the cap and stand upright like the ones on your motherboard now. Axial lead caps have leads on each end and lay flat - you can't use those. You probably won't even see any axial ones if you're looking at low ESR caps on eBay.

You can't go lower than 3300uf. You could go higher than 3300uf, but this would probably mean a bigger can size, too. There really shouldn't be any reason to consider other uf caps - 3300uf is a standard value and should be just as easy/hard to find as other values.

The Hong Kong sellers have great prices, but it usually takes a week and a half to get a package from them (shipping to the U.S.).

The only on-line component places I'm familiar with of are Mouser and DigiKey. They tend to be pretty pricey if you just need a few of something, and it's hard to figure out substitues if they don't have the exact component brand and value you need. I gave up after a few minutes on both sites. I'm sure they have something suitable, but I couldn't find anything obvious using their search filters.

I would think the badcaps guy or ebay are your best bet for these particular caps. Ultra-Low ESR caps this skinny with this much capacitance are relatively difficult to find, period. It seems like the people that do have them are selling them specifically for motherboards, so maybe that was about the only common use for them. That's about all I know about them. Maybe someone with more electronics knowledge can help out here.

[Tech_Cowboy]

Just an update to this thread.

I've replaced the leaking capacitors with some Sanyo WG 6.3V 3300UF Super Low ESR capacitors, as you can see in the image below:



This wasn't easy at all. The solder points on the underside of the board are very small. It requires a very fine tip soldering iron and some steady hands. It's easy to make a solder bridge to some other point if you're not careful.

But everything went well because I took my time doing it. And I'm happy to report that I can once again over-clock my CPU. The BIOS thinks that my Athlon 2500+ is an Athlon 3200+. The results from a few benchmark tests confirm about a 35% improvement in speed. It's been running stable for 3 days now. No unexpected shutdowns, no BSoD, no problems at all.

I'd like to believe that the caps were the reason that I couldn't over-clock before. But something else may have contributed to the problem: heat. While replacing the caps, I removed the heatsink from the CPU so that I had a little more room to solder. I noticed that the thermal pad or tape (not sure what it's called) between the CPU and heatsink looked kind of broken. Some of it had stuck to the heatsink and some had stuck to the CPU. When I tried to scrape it off the heatsink, it practically crumbled away. It had become dry and chalk-like. So I used isopropyl alcohol to get it completely off both parts, then applied some thermal grease (Thermaltake) and reassembled it to the board. Now, my CPU temperature is about 5°C cooler at 100% load, even though I'm over-clocked.

As I understand it, the A7N8X Deluxe will shut down if it determines the temperature is too high. So this could have been the problem all along. I guess I'll never know. But I'm glad I learned that lesson without frying my CPU. If I get a new processor and it comes with that pad or tape stuff, I'm going to immediately replace it with thermal compound. I just wish I had known this when I assembled my system four years ago. If anyone else is getting high temps, even with periodic cleaning, you might have the same problem under your heatsink.

Anyhow, that's the end of the lesson for today. I want to thank everyone who helped me out, especially Gibby. My system is speedy again, with the added benefit of being cooler.

TC

[Solarius]

Congrats, that's some hardcore repair you did there.

I would certainly echo your advice of always replacing the stock thermal compound upon purchase. I actually like to go in and reapply every year or two as well, as I find that (even though AC5 states it will NEVER deteriorate) it does dry and split a little over time. It's not like a tube of the stuff is more than a few dollars, and can cover 10-20 cpu changes (if you can get it to keep that long).

I would recommend that you purchase AC5 or creamique, or some other brand that specializes in grease, instead of using a tube that comes with some other part for free, especially if you're OCing. That being said, you'll probably be fine.

[PCBONEZ]

I want to straighten out some confusion about motherboard capacitor problems.

There are more than one cap problem going on.

Here are 4 problems (that I know of):

~~~

Yes there is the faulty P51 electrolyte issue.
(The famous "capacitor plague" that everyone knows about.)

The other problems are not in any way related to the P51 problem.

~~~

Nichicon HM and HN series capacitors manufactured between 2002 and 2004 had a manufacturing defect (Nichicon has given out two different stories about what the problem actually was.) - This is the one that gave Dell, HP, Intel, and MAC (G5's) all those problems.
-
I'm sorry I don't recall the exact months for the problem caps.
-
These dates are when the caps were manufactured and NOT when the motherboards were assembled. Motherboards may have been assembled with these caps late into 2005.

~~~

Currently the Chemicon KZG series is questionable.
(Nippon Chemicon, United Chemicon, same-same)
The KZG series is showing up failed with uncomfortable frequency. (Meaning often.) -
The interesting thing is there are usually no VISIBLE signs of a problem at all. (No swelling, no venting, no anything visible.) - Typically the motherboard works fine one day and then just refuses to boot the next day. Some people experience random shutdowns and BSOD for a while before it dies entirely.
-
Replacing just the KZG caps fixes the problem. (I have personally experienced this problem on (2) MSI boards and one each a Gigabyte and an Asus board. (All were P4 boards. All worked fine after replacing the only KZG's.)

This problem is inconsistent and Chemicon isn't saying anything so it is not known if this is a problem with all KZG series caps or only certain batches or sizes.

~~~

BIG BIG BIG.
This is a BIG one that gets confused with the P51 electrolyte plague because it affects capacitors manufactured in Taiwan.
BUT!
This one DID NOT GO AWAY and still exists.
-
A number of companies during 'the plague' swore up and down that they didn't use the P51, yet their caps were failing exactly the same way. - THIS IS WHY.
-
Caps that use Taiwan aluminum foil will fail the same way the P51 caps did.
~~~~
In 2004 the University of Maryland did a study to compare capacitors from Taiwan market to those from the Japanese market. - This study was initiated *because* of the P51 problems but they found another problem that IS HUGE and yet few people seem to know about it because they focus on the P51 issue and miss this.
-
(In addition to the P51 problem.) The study discovered that the aluminum foil manufactured for capacitors in Taiwan is below the accepted minimum purity requirements for aluminum electrolytic capacitors AND, in addition to simply not being pure enough, some of the impurities it contains are in high enough concentrations to cause chemical problems inside the capacitor that result in the same exact problems (bloating, venting, leaking) as the P51 electrolyte.
-
It causes THE SAME problem (gases forming inside the cap) as the P51 electrolyte did it just takes slightly longer to happen because the impurities have to 'leach' out of the aluminum before the chemical changes happen that cause the gas to form.
~~~~
BE ADVISED: Not all manufacturers in Taiwan use aluminum foil from Taiwan. Some always import it from Japan.

.

[MSIsucks]

Just thought I'd add my 2 cents. My Motherboard PCI express MSI 915P neo2 platinum came with 7 Nichicon HM(M) capacitors on it. All are bulging some vented thru the tops of the aluminum. It's not clear to my why MSI used these known bad capacitors on the board. Also, the PC no longer posts.

I guess now I know why the flagship msi grantsdale motherboard was so quickly swept under the rug. Literally a month after it came out it was removed from their website as current products now I know why.

[PCBONEZ]

The HM/HN issue wasn't know until sometime in 2004.
NOT expected. Nichicon is known for quality caps.

On the other hand MSI is NOTORIOUS for using crap caps on their boards.
(MSI was in fact sued for it over the 2000-2004 electrolyte issues and STILL continued using those caps during and after the law suit.)

If your board is newer than mid 2004 (by manufacture date) and your caps are 2002-2004 vintage then MSI knew better and used them anyway.

ECS and Soyo are the only major brands that are worse about crap caps than MSI.
There are 3 or 4 that may be running neck-neck with MSI for third place.

Contrary to popular belief Asus is hit-n-miss. They use questionable caps on about 1/2 their boards in the mid/low end price ranges. I have an Asus 'entry level' server board [Intel E7205 chipset] with crap TMS caps all over it. Asus got also caught with the HM/HN thing. (But as I said those caps were used by everyone for almost 2 years before the problem was identified.)

Abit and Intel are the only companies I have ever seen be proactive about replacing crap caps (that weren't known to be crap at the time they were manufactured) on their already sold boards. [Abit still got sued over it.]
-
All the others had to be sued to do anything.
(Or like Asus they got lucky and didn't get sued so they didn't do anything.)

Intel is the only one I know of that didn't get caught up in the bad electrolyte mess 2000-2004 at all. They were using primarily (probably exclusively) Japanese manufactured caps during that time.

I used to favor MSI and Asus but not anymore.
If I can't see a board in person to know what caps are on it I'll buy Intel boards.
Abit or Tyan are second choice unless the board was made between 2000 and 2004.

~
About 1/2 the above is just my opinion but consider that I inspect between 50 and 100 unique motherboards a year and I LOOK at the caps. (I'm trying to be retired but people won't let me... LOL)

~~

Also look INSIDE your PSU. A power supply with bad caps will blow otherwise good caps on a motherboard. Even Seasonic PSU's show up with crap brand caps now and then,,,, so look there too.

.