Abit KX7-333

[FFXIhealer]

Current Status: Working
Concern: small leaks and minor bulges on caps

Complete capacitor inventory and Color Code Legend:
2x Nichicon 4700 HD 6.3v H0147 BROWN
5x Nichicon 3300uF 6.3v (Tall Black near CPU)
2x Teapo 2200uF 6.3v GRAY
2 x Teapo 1000uF 16v YELLOW
14 x Teapo 1000uF 10v RED
8 x Teapo 1500uF 6.3v BLUE
8 x Teapo 1000uF 6.3v GREEN
2 x Teapo 470uF 6.3v PURPLE
5 x Teapo 100uF 16v CYAN



The ones with bulges and/or detected leaks are marked here in red:


If I'm going to bother with changing SOME caps, wouldn't it just be a good idea on a board this old (from 2001) to just get a total re-cap job and be done? The board is working as of this posting, but I plan to drop a newer "Barton" Athlon XP 3000+ processor to replace the current "Palomino" 1800+ chip and a much higher powered Radeon X700 Pro to replace the current Radeon 9550xl.

I have taken many pictures of this board's caps, but I only posted these two as the most informative. If you need other shots, I can post them. I hang out mostly at Vogons.org and a guy on Facebook clued me into taking a closer look at the caps (I had noticed the bulging, but not the leaking ones).

[Wester547]

Re: Abit KX7-333

I wouldn't be flabbergasted to find that all the Teapo are out-of-spec, bulging or not. So replacing all the caps on that 15+ year old board is a good call. I would guess the board has a mixture of Teapo SS, SM, and SC series, possibly one or the other for the latter two.

For the 3300uF 6.3V Nichicon HM, you can use these:

https://www.badcaps.net/store/produc...products_id=30

For the 4700uF 6.3V Nichicon HD, these are excellent replacements:

https://www.badcaps.net/store/produc...products_id=29

For the 1000uF 6.3V and 10V (if it's too tall for the 1000uF 6.3V, you can use the 820uF 6.3V ZLH instead):

https://www.badcaps.net/store/produc...products_id=46

For the 470uF 6.3V:

https://www.badcaps.net/store/produc...products_id=49

For the 1500uF and 2200uF 6.3V (if it's too tall for the 1500uF 6.3V, you can use the 1000uF 10V ZLH instead, and if you don't want to lower the capacitance on the other caps, you can use these):

https://www.badcaps.net/store/produc...products_id=39

For the 1000uF 16V (the height shouldn't be a problem, but if it is, you can use these):

https://www.badcaps.net/store/produc...roducts_id=156

For 100uF 16v:

https://www.badcaps.net/store/produc...roducts_id=175

[FFXIhealer]

Re: Abit KX7-333

I'm tempted to just send it in to be done by this guy, even though I am fully capable of doing the work myself, I only have a Radio Shack soldering iron, no workbench, etc. And I'd want all the caps replaced anyway and there's a good 48 caps. How to you get a-hold of this guy to do the work anyway? His Contact page gives a 404 error.

[Wester547]

Re: Abit KX7-333

Also forgot to mention that this is also a more precise replacement (capacitance wise) for the 470uF 6.3V capacitors: https://www.badcaps.net/store/produc...products_id=51

This page gives you a 404? https://www.badcaps.net/index.php?pageid=request
Here are the repair details: https://www.badcaps.net/index.php?pageid=repair_costs

If that iron is rated for 60W or thereabouts, it should be good enough for soldering on a motherboard (and certainly good enough for other projects involving multilayer boards). This board is probably old enough to use leaded solder, although the ground planes by the CPU VRM are thick enough to absorb surfeits of heat, thusly a more powerful iron is a necessity (or/and some flux, separate from the solder mix, ideally applied to the joints before soldering, as it allows for better heat transfer from the iron to the joints), but it will be easier to damage and lift traces with soldering irons of a higher wattage.

I don't think it would hurt to attempt the project yourself. But if you'd rather not do it with your current soldering equipment, the above URLs should allow you to send it to Topcat for repairing.

[FFXIhealer]

Re: Abit KX7-333

Also forgot to mention that this is also a more precise replacement (capacitance wise) for the 470uF 6.3V capacitors: https://www.badcaps.net/store/produc...products_id=51

This page gives you a 404? https://www.badcaps.net/index.php?pageid=request
Here are the repair details: https://www.badcaps.net/index.php?pageid=repair_costs

If that iron is rated for 60W or thereabouts, it should be good enough for soldering on a motherboard (and certainly good enough for other projects involving multilayer boards). This board is probably old enough to use leaded solder, although the ground planes by the CPU VRM are thick enough to absorb surfeits of heat, thusly a more powerful iron is a necessity (or/and some flux, separate from the solder mix, ideally applied to the joints before soldering, as it allows for better heat transfer from the iron to the joints), but it will be easier to damage and lift traces with soldering irons of a higher wattage.

I don't think it would hurt to attempt the project yourself. But if you'd rather not do it with your current soldering equipment, the above URLs should allow you to send it to Topcat for repairing.

Thank you for those links. They work. I've sent off for a quote. If the price is right, I don't mind providing a little business his way to get a professional job done. I'm comfortable soldering guitar parts together, but I really don't like the idea of accidentally messing up my PCB when I have a CPU and a video card coming in. And the site says he warranties his cap jobs for 2 years, so that's not bad either.

[Stefan Payne]

Re: Abit KX7-333

As an alternative you might also be able to use Polymer capacitors for the CPU VRM at least.

[Drack]

Re: Abit KX7-333

You can't blame teapo when nichicon have failed too

[Wester547]

Re: Abit KX7-333

^ Those Nichicon HM failed as a matter of course, because they are from a defective batch of HM (and HN) produced over the span of a few years (2001-2005). Those defective caps cost Dell a $300 million dollar charge for the 97% failure rate their computers (GX260, GX270, etc...) with those faulty HM and HN had (IE, millions of motherboards ceasing to work at the drop of a hat because of failed Nichicons in as little as a year's time). They were that bad. Honestly, I'm surprised those HM on that board lasted this long.

So yes, I do lay blame to the Teapo (not for the failed Nichicon as they are in different areas of the board - those Nichicon failed on their own). Teapo capacitors have a high failure rate, this is undeniably true as per the experience of the vast majority of members here, past and present, active and inactive, as well as other accounts on the internet of people dealing with prematurely failed Teapo capacitors.

[Stefan Payne]

Re: Abit KX7-333

^ Those Nichicon HM failed as a matter of course, because they are from a defective batch of HM (and HN) produced over the span of a few years (2001-2005).

And that makes the Teapo bad because of?!

First: Ultra Low ESR Caps are also rated for about 2000h of lifetime.
As are those Teapo SC, 1000-2000h the old ones, up to 3000h the newer ones. That board is from the beginning of the 2000s.
WHY should they be OK after 15 years?!

Those defective caps cost Dell a $300 million dollar charge for the 97% failure rate their computers (GX260, GX270, etc...) with those faulty HM and HN had (IE, millions of motherboards ceasing to work at the drop of a hat because of failed Nichicons in as little as a year's time). They were that bad. Honestly, I'm surprised those HM on that board lasted this long.

Capacitors fail, deal with it.
That's because of the chemical degredation due to chemical reactions (Water really likes to react with things!)

So yes, I do lay blame to the Teapo (not for the failed Nichicon as they are in different areas of the board - those Nichicon failed on their own). Teapo capacitors have a high failure rate, this is undeniably true as per the experience of the vast majority of members here, past and present, active and inactive, as well as other accounts on the internet of people dealing with prematurely failed Teapo capacitors.

If you have accusations, you have to prove it. Otherwise don't do it!

What you said here is just in the area of fairy tales, nothing more...

Because you have to take the specifications of the caps into account!!
Is the SC series a 'high reliability' one? No of course not! It's the buttom of the line Teapo has.
And YOU expect those things to last?!

And the failed nichicon you protect, but Teapo you flame?!
Pretty hypocritical, don't you think?!
Especially since we don't know how they were actually used, how hot they got, the ripple current they had to endure...

Yes Teapo SC might fail more frequently than Nippon KZM, but the KZM also has far superior specs, the lifetime is somewhere of 3 times that of the Teapo SC series. SO yes, the KZM might fail less than Teapo. Obvious with those specs, isn't it?!

[Wester547]

Re: Abit KX7-333

Okay, this will be my one and only reply to you. I'm not getting into a quarrel when this thread is about the OP's repair and not over a difference of opinion.

And that makes the Teapo bad because of?!

First: Ultra Low ESR Caps are also rated for about 2000h of lifetime.
As are those Teapo SC, 1000-2000h the old ones, up to 3000h the newer ones. That board is from the beginning of the 2000s.
WHY should they be OK after 15 years?!

They're rated at 105ºC. With accordance to the Arrhenius rule, for every 10ºC drop, their lifespan should be doubled, but that rule more or less refers to the rate at which the electrolyte evaporates by way of rubber seal diffusion or the self-healing mechanism, and has nothing to do with predicting what happens once the electrolytic chemistry falls asunder. It may also corroborate other factors, such as how much the electrolyte expands and contracts with heat and condensation, or leakage current characteristics, but generally references evaporation and nothing else. So using the rated endurance to predict the lifespan of a capacitor doesn't work and is a moot point. It doesn't explain how 2,000 hour @ 85ºC Japanese parts could still be good to this day but 105ºC rated parts of the same rating from Teapo could go bad, even in storage, because load life has nothing to do with shelf life, which is often much shorter (500 - 1,000 hours at the rated temperature). But even the shelf life cannot predict the numerously complex electrolytic chemistries.

Furthermore, both KZG and MBZ are rated for 2,000 hours at the rated temperature, but everyone on this forum knows that KZG (and KZJ) are so notoriously bad that they will die in a sealed bag, retaining their full leads. The same cannot be said of MBZ and MCZ. KZG's shelf life is rated for 1,000 hours at the rated temperature and KZJ's 500 hours at the rated temperature. MBZ and MCZ I don't know, but it's probably somewhere in that ball park. Again, you cannot use the datasheet's lifetime figures to estimate lifespan when it takes many more thousands of hours for unwanted chemical reactions and processes to occur (although heat will expedite that).

Capacitors fail, deal with it.

The issue isn't that liquid electrolytics fail. Each and every type of capacitor has their wear-out mechanism. It's the rate at which they fail, and for the most part, the Chinese and Taiwanese brands do fail faster than the Japanese brands.

That's because of the chemical degredation due to chemical reactions (Water really likes to react with things!)

Like I said in another thread, the Japanese majors have used the right additives, that is the necessary passivators, hydrogen scavengers, and corrosion inhibitors in the electrolyte to avert that reaction (hydration, foil corrosion, etc). The pH balance is also very important (a balanced concentration of hydrogen and hydroxide ions), which the Japanese majors get right. They also use ultra-pure formed, etched, cathodic and anodic aluminum foil (which oxidizes and etches much better than impurer foil, and impure cathode foil can release electrons which react with the hydrogen ions in the electrolyte to produce hydrogen gas, IE forming galvanic cells around the impurities), and all have a full, dedicated staff of QC chemists just for research and development, so they can devote much effort to fundamentally understanding and improving upon the process of manufacturing electrolytics.

If you have accusations, you have to prove it. Otherwise don't do it!

There are at least hundreds of posts on this forum dealing with prematurely failed Teapo, all the way back to 2003 (some of which show Teapo failing in unused equipment). All you have to do is search.

What you said here is just in the area of fairy tales, nothing more...

Fairy tales that other members will support.

Because you have to take the specifications of the caps into account!!
Is the SC series a 'high reliability' one? No of course not! It's the buttom of the line Teapo has.
And YOU expect those things to last?!

They're not failing because Teapo are being used outside their specifications. They're failing because the raw materials are crap and are degrading over time.

And the failed nichicon you protect, but Teapo you flame?!
Pretty hypocritical, don't you think?!

I wasn't protecting Nichicon or Chemi-con. I called them both out for their failures (though out of the thousands of series they've produced over time, only a handful have a disreputable history). But Nichicon stopped producing defective caps over a decade ago, and Chemi-con discontinued KZJ nearly a decade ago. I don't know if they ever fixed KZG but I know they don't produce ultra low impedance wet electrolytics anymore. Teapo, on the other hand, continue to make low-quality electrolytics to this day. I'm not chastising the fact that you don't mind Teapo. All the bad brands are inconsistent in quality so there are some members here who haven't had much of a problem with them. And whether they're "good" or "bad" all depends upon your definition of just such a term. For some, "good" is lasting a few years until they upgrade (or outlasting the warranty), which Teapo are capable of. For others, it's lasting a decade or so, which Teapo struggle to do outside of the voltage doubler.

Especially since we don't know how they were actually used, how hot they got, the ripple current they had to endure...

Well, it looks like the Teapo on this board were either directly filtering PSU outputs (not very stressful if a good PSU was used), additionally smoothing the output of linear regulators (not very stressful unless those linear regulators run insanely hot), or are filtering the input and output of buck regulators (more stressful but theoretically not as stressful as the CPU VRM). So yeah, some of them could be in potentially high ripple current, stressful locations. The Japanese equivalents (YXG, HE, ZL, etc...) might have dried up a bit by this point but certainly would not have bloated. While there's always some hydrogen escaping through the rubber bung and some vapor pressure inside the can (I.E, very little gas development), it should not be excessive (there should be depolarizers and hydrogen absorbers present in the electrolyte to prevent hydrogen atoms from bonding and forming molecules of gas), and so long as capacitors are used within their specifications, they shouldn't bulge under normal conditions, even after 15 years (unless they were running very hot).

Yes Teapo SC might fail more frequently than Nippon KZM, but the KZM also has far superior specs, the lifetime is somewhere of 3 times that of the Teapo SC series. SO yes, the KZM might fail less than Teapo. Obvious with those specs, isn't it?!

I wasn't talking about KZM.

[FFXIhealer]

Re: Abit KX7-333

Well, some additional info...

The motherboard was bought some time around the 2002 year and it was used until around 2005, when I joined the military and moved out. The PC then sat on a desk/shelf for years.

It wasn't until later in 2016 that I pulled it off the shelf and began messing around with it again. I put the system back in its original case and it did power on. Only now instead of 2 x 256MB DDR-266 sticks....it has 2 x 1GB DDR-400 sticks (which really did smooth out performance with Windows XP SP3).

The board still works to this day, although I will admit that it was a bit squirrely trying to get the two 1GB Memory Modules to play nice. Once the system booted with both, it's been stable since. Perhaps a symptom of the caps going bad after all this time?

And you two can argue all you want. I'm getting info about capacitors either way as I read them. Besides, it's fun to laugh at people bickering about trivial stuff on an on-line forum.

Honestly, I have more concern about the hard drive going bad than I was about the motherboard capacitors, but like I said before, someone pointed out the bulging and the small amount of corrosion on top of a few of them. And I'm the kind of guy who likes to prevent problems before they become an actual problem. I don't wait until my tires blow out on the side of the road - I replace the tires BEFORE they go completely bald. I change my oil between 3,500 and 5,000 miles.

And I'm still trying to decide whether to order the caps and do the work myself or to just send it in to the guy and know it'll be done right. It'll cost more and I'll have to wait longer, but there's something to be said about "peace of mind." Besides, I wouldn't have a clue where to start about replacing the caps with "better" ones, even switching out from electrolytic to solid-state ones or something. I know pretty much nothing about capacitors on computers. My knowledge of capacitors is limited to their use as high-pass filters in a guitar's tone circuit.

[Drack]

Re: Abit KX7-333

I said nichicon BECAUSE of the HD Failing, I know hm/HN and chemicon kzg/kzj fiasco as everyone in this forum, just saying that when hd fails you can't blame the teapos, I know they are not the best buuuuut they are not the worst you can get (capxon,lelon,sacon,fhy,samyoung,BH, fuhjkyouxD,asiacon,asiaX,I.o,zcky,tk,CA,jamicon...)

[Drack]

Re: Abit KX7-333

Well topcat has a guide of how to change caps, try it with an old dead board that's what I did, now I change caps like nothing, I think you can polimod after vrm where the 3300uf are we, need the datasets

[Wester547]

Re: Abit KX7-333

I said nichicon BECAUSE of the HD Failing, I know hm/HN and chemicon kzg/kzj fiasco as everyone in this forum, just saying that when hd fails you can't blame the teapos, I know they are not the best buuuuut they are not the worst you can get (capxon,lelon,sacon,fhy,samyoung,BH, fuhjkyouxD,asiacon,asiaX,I.o,zcky,tk,CA,jamicon...)

Sorry for the misunderstanding... but the HD aren't bulged on this board. Only the Teapo and Nichicon HM are (although I have always wondered how "safe" HD of that time period are). Personally, I would consider Samyoung to be better than most of those brands, but would certainly lump Samwha in right there with them. I don't know if the HD have gone bad without showing it or not, the OP just wanted to do a full recap to be safe after 15 years.

[Drack]

Re: Abit KX7-333

Forgot samwha Samsung should be ashamed

[FFXIhealer]

Re: Abit KX7-333

Well, I ordered a nearly-complete cap set for this MB. I left out the caps below 1000uF after talking about it with the guy over e-mail. I will be attempting this job myself. I'm crossing my fingers that I don't screw anything up. I have a 20-watt soldering iron and one of those 45+ watt ones. I think I might have a sucky-pump thing somewhere.

[FFXIhealer]

Re: Abit KX7-333

New Barton CPU works like a charm after I straightened out the crooked pins.

Wanna know what DOESN'T work? The X700 Pro. The system refuses to POST with it in there - not even a beep code. Just running fans and HDD whine (the HDD spins up, of course).

[Drack]

Re: Abit KX7-333

Might be a dead gfx, do you have another one to test the Mobo?
Did you recap it?

[FFXIhealer]

Re: Abit KX7-333

I'll answer that with a full UPDATE.

Received my new caps order today - found out that two of the caps I ordered have a larger diameter than the caps they're to replace, so now I have to adjust fire.

I attempted to de-solder the first of the tall 2200uF caps near the CPU....with a 20-watt Radio Shack iron and a 45-watt Radio Shack iron from the 80s or something. Neither one would melt the solder joint of the first wire, so I'm calling a halt to trying to do this myself. I just sent an e-mail to the admin guy here asking if I can still send the MB and all these caps to him for him to do and to just send me a corrected invoice for the labor/shipping and to swap out the caps I bought for the correct ones, etc.

The MB AGP slot works because I have a Radeon 9550 AGP 4x/8x card that works in this 2x/4x "Universal" (no key) AGP slot, so I know the slot at least works for the 9550 in 4x AGP mode. But the X700 Pro is a PCI-Express chip that is sitting on an AGP card, so there's probably some form of PCI-E -> AGP interface chip that's trying to lock the card in 8x AGP bus mode - which WOULD cause the whole system to lock up on POST. I even cleared the CMOS and tried it with basic defaults, and no-go. MAYBE it'll work after a recap job, but I doubt it at this point. I think it's an AGP 8x only card. Let's face it, the card's from 2004-2005 and all MBs then were AGP 8x boards or were switching over to PCI-Express. And 9800xt cards are hard to find and when someone does try to sell one on E-bay, the prices are high - like $50 minimum for a card that hasn't even been verified as working...just being sold as-is.

[FFXIhealer]

Re: Abit KX7-333

MB shipped out today with the caps - well packaged with the instructions manual too. LOL Here's hoping that the end of February brings to me some solid retro gaming.