Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

[Ginguskhan]

Hi fellow Badcaps forum members!

First, thanks for taking the time to read this over and maybe even offer me some advice.

My System (still running stable, despite 3 bad caps)
Mobo: MSI K8N Neo-4 Platnium (7125. PCBv1, Bios version: 1.B) Mobo specific link: My MOBO link
CPU: FX-55 on stock cooler
PS: Ultra 500w ATX (checked, no bad caps)
1 GB pc-3200 Corsair ram, 320 Seagate HDD, Radeon x300SE

Right next to CPU my mobo has 1 leaking cap (circled in red in pics) and 2 vented caps (circled in yellow in pics):
All labeled the same:
KZG - 3300mF - 6.3v
Back: (M) 105 C<br>5(2) 3P

Also near these 5 caps and the cpu is a cap (EC26) labeled:
KZL (where the "L" is backwards?) 1800mF 6.3v
Back: (M) 105 C<br>A(3) 53

Please see my pics, hopefully they are informative.

I found a very similar post in badcaps MSI forums, perhaps this is right: MSI K9N Mobo

Recap:

ARACH Posts: "I have a MSI K9N Platiunum. I'm familiar wIth the cap issue and was an electronic technician. I was experiencing random shut downs and BSOD's. Opened up my box and sure enough I have three bad caps right between the CPU heatsink and the power supply tranistors heatsink. There are six total. They are 3300mF 6.3v. The brand is KZG. Has anyone heard of this brand or issues with this particular board."

Leopard suggests: "Those are Nippon Chemi-Con KZG caps. Chemi-Con caps are usually good but KZG series is prone to fail early under high heat stress. Replace all KZGs near the CPU."

PCBONEZ: "The above is correct.
Chemicon KZG are in common use and have problems near heat.
- You're lucky in a way because they don't always bloat when they fail.
Bad caps in the PSU might have helped them along.

Specs wise they (in 3300uF/6.3v) are rated at 2800 for Ripple and .012 for ESR.
That's on a par with Rubycon MBZ, Sanyo WG, Samxon GD, and Panasonic FJ (in 10x25mm for FJ). "



Questions:


Cap replacement questions (replace 3,5, or 6 caps?)

1) There are 5 identical caps near the heatsink/cpu (3 of which have signs of problems)...should I just replace all of these, even the ones that appear visually ok?

2) What about the random cap (EC26) that is 1800mF/6.3v KZL(backwards "l"), should I replace this one? If yes, what cap should i use as a replacement?

3) Among the caps PCBONES mentions, I have access to Rubycon MBZ, Sanyo WG, and Panasonic FJ (1800mF/6.3v). What is the best choice out of these? Note, I can get 5x sanyo WG for 4.50 shipped, is this a good deal, and it will work ok?

4) For good cap brands, do I need to measure the ESR/capacitance of new/factory sealed capacitors or is it safe to assume they are ok? Can a capacitor age when sitting in inventory, factory sealed for say 5 years?

5) Random: Has anyone had luck flashing MSI MOBOs with the "MSI Live Bios Update"? Or do you stick with the floppy dos based approach (I dont have a floppy drive)


Thanks,
Matt

[gdement]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

1) There are 5 identical caps near the heatsink/cpu (3 of which have signs of problems)...should I just replace all of these, even the ones that appear visually ok?

You should replace all of those. The ones that haven't failed yet, eventually will. They're the same caps and probably connected in parallel on the same circuit. If any of them failed, they should all be replaced.

2) What about the random cap (EC26) that is 1800mF/6.3v KZL(backwards "l"), should I replace this one? If yes, what cap should i use as a replacement?

That's a KZJ.
These don't have an established pattern of failure, but the manufacturer advertises their relationship to the KZG series. So by association some are uncomfortable with KZJ and would say to replace it.
KZJ is one tier above the KZG. They're on par with Rubycon MCZ, Nichicon HN, Samxon GA. Not sure if there's a Panasonic equivalent, maybe FL.
If you replace with a Nichicon, make sure it's dated from at least 2005 or newer. Older ones were defective.

3) Among the caps PCBONES mentions, I have access to Rubycon MBZ, Sanyo WG, and Panasonic FJ (1800mF/6.3v). What is the best choice out of these? Note, I can get 5x sanyo WG for 4.50 shipped, is this a good deal, and it will work ok?

The WG would be fine. All the caps you mentioned are good to replace the KZG's, I don't know of any advantage between them.
The FJ 1800uF 6.3v doesn't quite meet the spec of the cap it's meant to replace (KZJ). But in practice it would probably be fine.

4) For good cap brands, do I need to measure the ESR/capacitance of new/factory sealed capacitors or is it safe to assume they are ok? Can a capacitor age when sitting in inventory, factory sealed for say 5 years?

In principle they do age after a few years. In practice it doesn't seem to be as much an issue as the manufacturers make it out to be. I've used plenty of 4-6 yr old caps on motherboards and they work fine.

5) Random: Has anyone had luck flashing MSI MOBOs with the "MSI Live Bios Update"? Or do you stick with the floppy dos based approach (I dont have a floppy drive)

I used the live update on my nephew's 945 Neo5 board. I'm not sure if it worked or not. The BIOS was updated, and gave the correct date at boot, but it didn't add support for a processor that supposedly was added by the update.
MSI themselves had 2 different versions of their CPU support chart which contradicted about this processor, so I don't know what to make of it. I never tried reflashing with the floppy.

[Ginguskhan]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

**UPDATE!**

Thanks gDement! And PCBONES!

I finally got my caps in the mail today....and success! I got 5x Sanyo WG 3300uf, and when I measured them they all were right around 3500 to 3600uf on the cap meter I was using (all measurements seemed kind of high).

First, I measured the caps on the board:

All 6 caps near the cpu = 5000uf (must be in some series/parallel circuit)

Once I got the soldering iron temp right, all 5 caps fell out pretty fast.

After removing them from MOBO my old caps measured:

Oozing cap = 700uf --> not good
vented cap EC21 =1580uf
vented cap EC17 =1000uf --> how was my MOBO still stable?!
EC27 = 4000 uf
EC29 = 3800 uf
**EC26 = 1790uf (this was the measurement once i removed 5x caps, 1 more sig fig since I could use 2000uf cutoff on cap meter)

I'm going to keep an eye on EC26, but it tested ok once I isolated it from its original circuit with the 5x 3300uf caps.

Clearing the cap holes was difficult, even with a professional solder remover. Took me a couple tries on a few holes, but I obeyed the Golden Rule and took my time. Finally got it done!

The final step of soldering the new caps in was a breeze, lol it took me 1 min per cap! I recommend using SN60/PB40 solder or something close to it (I had 63/37).

Total work time = 3.5 hours.....with me referencing the website word for word, triple checking polarities and cap measurements.

I also was lucky enough to have a friend who allowed me to access professional grade equipment at an engineering lab at my university

I may be a "nub" at this, but......
Believe in yourself! You can do it with the right answers, the right tools and taking your time!

Badcaps.net saved my MOBO! Thanks again guys!

[Ginguskhan]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

Greetings again Bad-caps forum readers!

This is the 2nd recap of my MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B). Since the first recap, it has been working a-OK with more stable vCore voltages! This second installment is a preventative recap; it eliminates all potential capacitors that could go bad (i.e. KZG/KZJs) in the most demanding circuit in a MOBO: the +12v VRM Input and VRM output (vCore). See pic for suspect KZG in +12v VRM Input circuit.

Capacitors in the +12v VRM and vCore experience more sustained heat, current and ripples vs. other MOBO circuits; thus, it is a good idea to have good caps in these two circuits. Aside from the peace of mind of having Sanyo WGs and Rubicons stabilizing my VRM voltages and consequently my cpu vCore for the long haul, aesthetically those green WGs and the midnight black MCZ look nicer than the poo-stain brown NCC KZGs they are replacing.

Re-cap #2 Gameplan:
1) Remove and measure 4x 16v 1500uF KZG (+12v VRM Input) + 1x 6.3v 1800uF KZJ (vCore)
2) Measure and replace KZG in +12v VRM w/ 4x 16v 1500uF Sanyo WGs
3) Measure and replace KZJ in vCore circuit w/ 1x 6.3v 1800uF Rubicon MCZ
Note: Polarity conventions on this board are standard (white is (-)negative/ground)

Execution:
1) Remove and measure 4x 1500uF KZG (+12v VRM Input) + 1x 1800uF KZJ (vCore)
I set my solder iron to 600C, primed it with some fresh solder and heated the old solder joint of the appropriate capacitors from the back of the board. I would heat up one leg at a time from the backside of the board, pulling softly once the cap felt loose. Basically, I rocked the old cap back and forth while still making progress toward pulling it out of the board. Usually the cap had made enough progress after 4-6 heatings (2-3 on each leg) to be freed from the MOBO.

Removed caps measured:
4x 1500uF KZG (+12v VRM Input) = circuit before removal = 5750uF, Individually after removal from board (see pic): 1430uF, 1470uF, 1460uF, 1460uF --> KZGs SHOULD NOT BE USED BY MANUFACTURERS IN +12v VRM INPUT OR VCORE!
1x 1800uF KZJ (vCore) = circuit before removal = 16900uF, after removal = 15530uF, Cap measured after removal from board: 1810uF Guess KZJ might be legit after all!

2) Measure and replace 4x 1500uF KZG in +12v VRM Input w/ Sanyo WGs These holes were large, so clearing them with the solder sucker worked on the first try @ 700C. I had to wait 20-45 seconds depending on how good a contact I had between the solder sucker tip and the solder on the backside of the board before pulling the trigger to ensure a clear hole. Caps slid right in, bent the legs to hold the cap in place. Then, I flipped the board over and soldered them in in a few minutes (used 63/37 solder)

3) Measure and replace 1x 1800uF KZJ in vCore circuit w/ Rubicon MCZ
This hole was so small I could not clear it after 2 attempts with the solder remover. So, I used a new technique: I heated the solder on the backside of the board, and held the new rubicon MCZ capacitor on the otherside of the board. Once the solder heated to a molten state through the length of the board (15-45 sec), one leg of the capacitor slid through with minimal pressure (Slide longer positive pole through board first, then shorter leg)! Then, I heated up the other leg's solder joint and the other leg went through....essentially just the reverse of cap removal I described earlier!

Final Cap Circuit measurements:

4x 1500uF KZG (+12v VRM Input) = circuit before removal = 5750uF
After 4x 1500uF Sanyo WG replacement 6250uF

1x 1800uF KZJ (vCore) = circuit before removal = 16900uF
After removal of KZJ = 15530uF
After MCZ replacement = 16800uF Guess it measures the same with both caps in place, but Id rather have an MCZ Rubicon stabilizing my vCore voltages over the long run!

KZG kill recap/count this board = 9 (+rep)
KZG recap/kill count totals = 15
Total recap/kill count totals = 16 recaps!

[Ginguskhan]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

Final Notes and Thoughts:

-Polarity conventions on this board are standard: white is (-)negative/ground
-Use 60/40 solder or similar (I used 63/37 solder = 63% Tin and 37% Lead)
-Tip: Cut small portions of electrical tape and cover the board surface near where you will be soldering. The tape gives the backside of the board and traces a second or two of extra protection from the sharp and scalding hot 600C solder iron tip. Also, It looks professional at the end of your recap when you peel away the tape and see ultra clean board, vs. board with some thermal damage, some solder or solder resin deposits.

[yonu]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

hey not to hijack your thread but I am working on the same model board and it had the same bad caps you showed so i replaced those however i am at a loss as the issue I am working on now is as soon as power is applied to the board it turns on however it doesn't post and I am not sure which capos to check as the rest of them are below 1000uf.

any ideas you might have would be great.

Thanks,
yonu

[PCBONEZ]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

KZG are known to fail with no visible signs at all.
I've seen it -personally- several times.
There have been enough reports about it now from various people to say it's a typical failure mode for KZG.

KZJ supposedly ['supposedly' by way of a blurb in an ad that used to be on their site] use the same electrolyte as KZG and most likely have the same problem. Thing is KZJ aren't widely used [not too many 'out there'] and since there are few to start with there are few reports.
.

[yonu]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

simple question d I replace the caps that are at 1000uf and below or are those normally good also can i replace the 1000uf one with 1500 or does it need to be 1000uf?

[PCBONEZ]

Re: Question MSI K8N Neo-4 Platnium (7125.v1, Bios: 1.B): 3 vented caps, HELP!

Not really that simple.
5-6 years ago replacing down to 1000uF was the standard thing.
About 4?? years ago it got lowered to 470uF by many.
Now we are finding some specific boards with blown 100uF and 220uF and in some cases even 22uF caps. [Usually due to cases with bad cooling or the use of 85c caps, or both.]
-
This will cover you 99+% of the time.
--> Of the bad caps makes or series'....
Replace all that are 8mm and up.
and
Replace all that are 470uF and up.
and
Replace all that have vents. [Sometimes 6mm caps have vents.]

On a mobo replacing one 1000uF with one 1500uF will -probably- be fine.
-
Replacing [say] ten+ 1000uF with ten+ 1500uF -might- be another story.
- Hard to say.
More uF means it takes longer for the voltage to come up because there is more capacitance to charge. Also means it will take longer to power down when you shut it off.
. On the bad side: that -might- screw up the power sequencing. If the rails on the board aren't at [some] % of the rated voltage within [I think it's] 50 msec then the chipset shuts the PSU down. That's a protective feature that's built into the mobo.
. On the good side: you will be slightly more resistant to brown-outs.
-
It's hard to say for sure because there is no way to know if the [in your example] 1000uF are on the high or low side of the acceptable range for any given board.
-

I did something similar replacing ~10? 330uF with 1000uF on an old P3 board as an experiment.
Works fine except when I power down it takes FOR EVER for the +5vsb light on the board to go out. [Don't wanna be unplugging RAM or add-in cards until that light is out!]
.