P4c800-e

[TBoneit]

Well I started looking closer at that Asus motherboard and it is mostly Rubycon & some KZE although the age could mean that the Rubycon are from the bad era.

It had Seven OST next to the CPU when I looked closer so I changed them out, said to myself there are nine places for them so I decided to put 9 1500uf & 6.3v in and then found out why the two spots were empty. They were in the way of one corner of the HSF assembly latch. My suspicion is that originally they were going to use nine 1200uf @ 6.3v. A slight design oops.

I took the closer look as it seemed OK one day and the next day I went to reload windows onto a 200Gb SATA instead of the little test drive and blammo problems again.

All of what I know about modern electroylitics and their use in modern eclectronics I've learned here.

Many thanks to all those who share their expertise.

RT

[PCBONEZ]

Re: P4c800-e

Jap caps didn't have a bad era.
Nichicon had a bad run of two Series [HM and HN] from 2001-2004.
Chemicon KZG and KZJ have high failure rates but dates don't matter so it's not an era.

OST aren't Jap and will fail without bloating no matter when they were made.
[Very similar failure to the KZG and KZJ Chemicons.]

The 'Capacitor Plague' is actually over 1/2 dozen issues with different causes [some of which have existed for decades and some of which are current problems] so when some web page says "THIS was the cause" or suggests the cap plague went away they don't know what they are talking about.
-
Wikipedia is about the WORST source of info on Cap Plague. I've corrected it a dozen times at this point [over the last 4-5 years] just to have some twit that knows everything about 1/10th of the problem and nothing about 90% of it to come in afterwords and screw it up again.

Also [This one is VERY common about the web.]
When you see someone talking about the Dell issues and blaming it on the 'stolen electrolyte formula' you immediately know the author doesn't know his/her ass from a hole in the ground.
- The affected Dells used 100% Jap caps.
- The 'stolen electrolyte formula' only affected CH and TW made caps.
How can caps that were NEVER USED in those Dells be the cause of their problems?
- DUH!
.
.

[seanc]

Re: P4c800-e

You mention this 200GB SATA drive again.
Since your small test drive worked ok, try another *known working* SATA drive please.

[PCBONEZ]

Re: P4c800-e

Some kinds of PC's blow even Jap caps regularly but that is a Heat Issue [aka inadequate cooling and poor case design] and not due to the caps.
The most common place to see this is Dell Optiplex mini-cases but those are FAR FAR from being the only systems with that problem.
It is simply seen most often in Dells because Dell makes a Cubic-Butt-Ton of Optiplexes and most others are lower in qty of production.
-- Dell Optiplexs -also- had the HM/HN issue from 2001-2004.
-- That means for a while they had *TWO UNRELATED* problems at the same time. [Heat and bad HM/HN.]
-- Neither of those problems had to do with CH/TW caps or the stolen formula.

Also just in general.
The HM/HN issue is OFTEN confused with/blamed on the the stolen formula fiasco because the two events occured at approximately the same time.
Nichicon makes their own electrolyte in Japan.
Nichicon NEVER used the electrolyte that was made with the bad formula.
The bad formula was made and sold in TW to companies that buy [vice make] their electrolytes.
-
The bad Nichicons showed up in virtually every brand of motherboard [including MAC] because they were considered high quality caps.
[And barring the defective run, they ARE high quality caps.]
The mobo manufacturers THOUGHT they were making their product MORE reliable by using [or changing to] the HM & HN.

For some mobo manufactures the timing lead to the following scenario where they TRIED to do good, but got screwed.
[Abit & Gigabyte bit it big time this way.]
- Using CH & TW caps and 'stolen formula issue' becomes known in 2001/2002.
- Switch to all Jap caps to avoid bad caps on their mobos.
- They chose to use Nichicon HM and HN - which were defective since 2001 - but that wasn't known until 2004.
All the customers knew was the boards were still blowing caps. They assumed no effort was taken to correct the problem.

Then there is a third caps issue that gets confused with both the HM/HN problem and the Bad Formula problem.
The Aluminum sold in CH & TW for cap foils [and cans] is less pure and has other impurities than Jap made foils.
Those impurities lead to the electrolyte breaking down causing cap failure.
This affects ALL caps using Aluminum foils made in CH or TW.
- This problem has existed longer than I have and it's a current on-going problem.

- And now you know why so many of us here stress using only Japanese caps as replacements.

Another little aside:
All the 'bad formula' caps also have the 'bad Aluminum'.
But only some of the 'bad Aluminum' caps have the 'bad formula'.

Visually the failures from those three problems are the same but internally the root cause is entirely different.
.

[brethin]

Re: P4c800-e

If it worked before you recapped it and added the 2 extra caps then my guess is you should have left it as is. Never just add a cap to a empty spot just because it will take one unless you know the circuit and know how it will effect the circuit.

[PCBONEZ]

Re: P4c800-e

My suspicion is that originally they were going to use nine 1200uf @ 6.3v. A slight design oops.

No. [Two reasons.]

1:
There is a minimum uF needed in the Vcore side of a VRM but:
- [B] the minimum required total uF is more like 5000-6000uF.
- [A] uF is not the most critical 'spec' there for caps, the ESR [meaning total] is far far more important.
-
7x 1500uF [10,500uF] is well above the minimum uF they needed, yet it's a pretty typical value on skt 478 and skt 775 boards with lytics.
So why do they do that:
- Higher uF caps have larger cans.
- [Within a grade of cap] The larger the can is the lower the ESR is.
They use so many [or so large of] caps there to lower ESR, not raise uF.

2:
When they design a board they often include places for 'extra parts' in the layout.
- One reason is to allow different versions of the board. [Like say with or without a SCSI controller or sound or fire-wire or additional USB ports] This allows them to use one board layout for multiple models.
- Another reason [when it comes to caps] is they aren't confident in the design's stability when it's just 'on paper' so they add a few extra cap 'spots' which [after testing an actual physical board] may later be determined to be unneeded. Since it costs money & time to go back and remove those spots from the layout, the silk screens and programs in their CNC drilling machines they just leave them there.
- A third reason [common near CPU sockets] is they want an upgrade path for the board layout should a CPU be released that is more demanding. This way rather than designing a new board from the ground-up they can re-use the old layout by changing the parts they put on it and have a 'new model' for the new CPU in production more quickly. [This senario happened a lot when skt 478 Prescott was released. Kept the old layouts but added some caps and changed the Chipset, MOSFETs, a few controller IC's, etc that they soldered onto the same old "blank boards".]
- And the most guessed [but least probable] reason is to allow using more 'low ESR' [cheaper] caps vice fewer 'ultra-low ESR' [mo $$] That made sense about 8-10 years ago but the cost difference isn't that much anymore. In fact it might even cost more to use more but cheaper caps now.
.

If those two caps you added are in parallel with the other 7 then it should not have caused a problem at all.

Did you check to make sure they ARE in parallel?? - With a meter??
Because -where- a cap is isn't a reliable indication of that one way or the other.
.