Replacing caps: how close does it have to be?

[willawake]

Re: Replacing caps: how close does it have to be?



the four near the printer port right? what brand are they on your board?

they are for 12v input not output of vcore (which will be the 6.3v ones next to the cpu)

what brands of caps are next to cpu on your board?

if you change 4 x 1500uF 16v with 4 x 1200uF 16v it is not 300 it is 1200 total less. this is almost 1 cap less total. it may be too much or it may be ok. it will not damage anything though but it will probably not be an improvement.

you are taking caps from a dead board? maybe the caps on that board are in bad condition, it is not recommended.

[tux_topo]

Re: Replacing caps: how close does it have to be?

Yes, the ones near the printer port. They are marqued with a M.
The dead board, died for a bad installed memory, wich burned out the chipset (please, don't ask how it could be, because i still trying to figure out).
BTW, where did you get the image?
PS: please, forgive my english, but i'm from Argentina, and it's easy read than write .

[Elitist]

Re: Replacing caps: how close does it have to be?

Personally, I wouldn't hesitate to make your proposed substitution. The 1200's may already exceed the 1500 nom. if you take into account the +50% tolerance. If they're old and not bulged, that's the best portent for the future you could have. Incidentally, those four are not directly paralleled - they act semi-independently for each regulator. Contrary to all the theory and specs, in practice, I've found old, quality caps to be more effective and longer lived than many newcomers! Notwithstanding, punch-through can occur at any time with any electrolytic cap. - nature of the beast, I'm afraid. Best option is to strip down some surplus mil. or space electronics if integrity is paramount. You can't buy quality like that!

[willawake]

Re: Replacing caps: how close does it have to be?

They are marqued with a M.

need more info, tell me everything written on the caps.

also the ones near the cpu

pic is from extremetech

[willawake]

Re: Replacing caps: how close does it have to be?

one thing i wanted to work out was whether tux_topo already had good caps on his board and perhaps there is a design problem instead. the review sample has chemicon kzg which i would consider to be ok.

it would be good to compare the caps on the board with the ones from the abit to see if it would be an improvement or not.

[tux_topo]

Re: Replacing caps: how close does it have to be?

OK... let's go to parts. I need to recap a total o 9 caps. 5 x 6.3v 3300uF, which apparently i have a not good photo of it, will be replaced by Rubycon MBZ with same values. The others 4, 1500uF 12v, y made a draw of the simbol, because i'm unable to get a decent pict to of it, that will be replaced by a blinded capacitor of 1200uF 12v...

The m:


And this is the pic of the 3300uF cap (it's the blurry one )

[willawake]

Re: Replacing caps: how close does it have to be?

on that board the 1500 16v are chemicon KZG, the M caps are probably panasonic FM. there is a rubycon and also some minor rubycon ZL. they are all very good.

i dont have specs on MBZ but KZG is a good spec cap, probably similar or better than MBZ. i dont see reason to replace the caps on your board. they are very good.

what caps are next to the cpu socket? show me a pic

[tux_topo]

Re: Replacing caps: how close does it have to be?

The pict that i show you, supouse to be for let you know about the caps near the cpu socket, you can read the 3300uF and 6.3v on it... And the 1500 16v that you can see at the top right in the image, are the ones labeled with the "M".
Hope to clarify something (because i didn't understant miself at all )

[willawake]

Re: Replacing caps: how close does it have to be?

ok i mixed up the FM and KZG

what i said is that the caps on your board are very nice. there is no reason to change them.

what is your problem really?

your vcore is unstable, how much does it fluctuate? do you have crashes? if you measure vcore with bios it can be inaccurate.

what psu are you using?

[tux_topo]

Re: Replacing caps: how close does it have to be?

The PSU is a Thermaltake Purepower Butterfly 480w. I don't have crashes, but the vcore fluctuates almost .05 all the time :s. And i test the rest of the pc, with a DFI LanParty nf4 sli DR, and be able to get a very stable vcore, and overclock. For example, the MSI at 1.5v vcore, a only can get 245 of fsb at multi of 9 and very unestable, but in the DFI i could achieve 280 x 9 at 1.475, rock solid.

[Elitist]

Re: Replacing caps: how close does it have to be?

You need to log your suspicious board with a storage scope (rather than a dc logger, which would be heavily damped and of poor response characteristics). Look at long term and short term excursions under working conditions as well as quiescent.
If you can't borrow a scope from the lab, you can purchase an adequate one for ~£200 and it'll probably pay for itself if you intend to do much troubleshooting and repair.
Notwithstanding, it seems we may be confusing different attributes of the boards here. What exactly is being measured, at what location and with what instrument?

[willawake]

Re: Replacing caps: how close does it have to be?

the dfi is obviously a better designed board. to improve the msi its going to be more job than changing a few caps. https://www.badcaps.net/forum/showth...5&page=7&pp=15

[Elitist]

Re: Replacing caps: how close does it have to be?

He seems to be confusing instability due to o'cing with a PSU that tries to follow that instability. Probably there is nothing wrong at all - as you say, the MSI is just a cheaper board!

[tux_topo]

Re: Replacing caps: how close does it have to be?

OKay... And the fluctuation in the vcore, y have all the time :S even if i'm in stock. And don't have a adequate method to measured the voltaje in that board :$. I only can use the bios .
Thanks anyway... i will gone a try to change mine board...

[Elitist]

Re: Replacing caps: how close does it have to be?

In which case, what you see is entirely spurious and can be safely ignored.