Ms-7093 ver:1.0

[ForumNut]

I am working on an older E-Machine T6212 with a MS_7093 MB.

It has at least one swollen cap. right next to the on board video.

I would like to identify the manufacturer and replace it with a better quality part.

Identifying markings:
OST
6.3v
1000uF
RLX
O503
105 C

I posted an image. I will post better images when the board is out of the machine tomorrow. The colors are blue with gold print.

Requesting any additional information on this issue.

Thanks

[steve2]

Re: Ms-7093 ver:1.0

http://www.ost.com.tw/products_ec_competitive.asp has rubycon mbz and panasonic fj as replacements for the RLX series. The are more, but I only checked the rubys and pannys.

[PCBONEZ]

Re: Ms-7093 ver:1.0

You should replace at least all the OST that are 6mm and over in diameter.
OST have a habit of going bad without bloating and as one has obviously gone south the rest aren't far behind [or have in stealth already].
.

[ForumNut]

Re: Ms-7093 ver:1.0

Here are the additional photos. I hope these are much clearer.

Thanks guys I will check out those sites.

Is there a brick and mortar store in the midwest USA that sells a good soldering iron for this job?

The only ones I have are not good for fine work like this.

What would be the best replacements? I see that OST doesn't have a great rep.

Who makes the absolute best Caps?

[PCBONEZ]

Re: Ms-7093 ver:1.0

The replacements will depend on what the originals are.
Are all of the OST on it RLX series?

Do you have a Fry's Electronics?
.

[PCBONEZ]

Re: Ms-7093 ver:1.0

It looks like you have Chemicon KZG in the VRM/Vcore.
Those [and KZJ] have the same issues as OST and should be handled the same way.
[The other series of Chemicon are fine. Only KZG & KZJ have problems.]

[PCBONEZ]

Re: Ms-7093 ver:1.0

You got the full treatment on that one!
I see a Nichicon HM and on another Nichicon I see a 2004 date code.
Nichicon HN and HM series were defective from 2001-2004 due to a manufacturing error.
As with the Chemicon those are the only 2 bad series with that brand.
[And only those with 2001-2004 date codes are bad.]

You also have some Teapo scattered about but they appear to be smaller than 8mm and the small ones don't fail that often.

RLX, HM, and KZG have comparable specs.
They can be replaced with:
MBZ [rubycon], GD [samxon], WG [sanyo/suncon], FL [panasonic], FJ panasonic, HM w/good dates [nichicon].
Or these which are one grade better:
MCZ [rubycon], HN w/good dates [nichicon], GC [samxon]

..
If you would like walked through HOW to figure that out, I can do that too.
.

[ForumNut]

Re: Ms-7093 ver:1.0

I must have gotten lucky then because this is the very first time I have had to do anything at all to this machine. Its been smokin' fast and has never even needed a windows recovery. It's about 5-6 years old I think. And it still seams to be able to most anything I try to get it to do. With a good video card I can even play Crysis. It cost me $250.00 as a refurbished box, and I never looked back. I also have a gaming machine that I built with a Nvidia 780i SLI and it's not really noticeably faster until you start stacking crate or barrels in Garys mod or something like that. And I am pretty sure that the SLI video cards are responsible for that performance.

I'm sorry to hear that I got so many poor quality components. But I will remedy that very soon I hope.

With the right equipment I am sure I can do this with very little problems. My biggest issue will be finding the proper components. I have already tried Radio Shack, Frys Electronics and most other brick and mortar stores in the area that I can think of. No one seams to have the proper parts. In fact I cant even find the one 6.3V 1000uF cap that is obviously bad...I think your right about replacing them all but I think I want to start with just the circuit that is a problem. I am experimental like that. Then once I get a benchmark, fix the rest to see if there is an improvement in performance. I don't think I will see any stability difference then the stock built board because its NEVER had any kind of stability issue up until now. But I want to have all the parts on hand to start and finish the job as I play with it.

FN
_______________________________________________

[PCBONEZ]

Re: Ms-7093 ver:1.0

Brick and Mortar stores that carry Low ESR caps are rare.
I don't even know of one.

Best you can do is build a complete list and order on-line.

This site [BCN] sells caps, the other popular places are Mouser and Digikey.
https://www.badcaps.net/
Digikey.com does not carry the needed grade of cap.
Mouser.com only has HM and HN in the needed range.
There is also BDent.com who carries mostly Sanyo/Suncon as far as low ESR goes.
[BDents part search is rather sucky and hard to use though.]

.

[ForumNut]

Re: Ms-7093 ver:1.0

I am assuming that the best components for this project are the MCZ [rubycon]s Is that correct?

[PCBONEZ]

Re: Ms-7093 ver:1.0

If you can find them.
MBZ and MCZ are now 'obsoleted' and out of prduction.
When they're gone they're gone.

As I said in the earlier post:

MCZ = HN = GC
... and they are 1 grade better than the original caps

These are the original grade and would be fine.
[RLX = HM = KZG] = MBZ = GD = WG = FL = FJ = HM
.

[ForumNut]

Re: Ms-7093 ver:1.0

I guess I am just unfamiliar with the grading system. Can you give me a quick lesson? As in what do these grads mean ect. Is there any that are 2 grades better?

Is Nichicon the next best option to Rubycon, followed by Samxon? Or all these the same grade? I guess I am just a little confused as to what makes a cap a grade higher. I think as long as I have to do this anyway, I may as well go all out as long as it's cost appropriate.

If it was your machine and assuming you wanted to keep it, what parts would you put in it?

[PCBONEZ]

Re: Ms-7093 ver:1.0

There is no official grade system but a term is needed to describe the situation.
If you put all the caps specs into a chart by ESR/Ripple and can size as I [and some others] have it becomes clear there are steps.
[uF and volts don't affect ESR and Ripple, the can size does.]

= is an equals sign
They are equivalent for ESR/Ripple when the can size is the same.

>>> If it was your machine and assuming you wanted to keep it, what parts would you put in it? <<<
MBZ, WG, FJ, HM
[Unless I just happened to have MCZ or HN left over from another project.]

Nichicon HZ and Samxon GA are one up from MCZ, HN, GC.
After that you would have to go with polymers.
... And I [me. myself] would if I wanted ESR that low.

IMHO: [and it is -just- my opinion]
HZ & GA are kind of pushing the limits in so far as durability/stability of a liquid electrolyte. I don't really trust them to last long and as they are relatively expensive they are rather rare 'in the wild'. So, despite the fact they aren't new series there isn't much of a track record to look at.
.

[ForumNut]

Re: Ms-7093 ver:1.0

Looks like I have to take a physical inventory of my board because I can't seam to find a schematic diagram to pull the cap values off of...

Should I get higher voltage values?

If I am understanding this right I'm looking for:

Quality Cap Manufacturer with

1. low ESR
2. Low Ripple
3. Higher Voltage
4. Exact uF values specified by the board manufacturer
5. The same size can.

Is that correct?

[PCBONEZ]

Re: Ms-7093 ver:1.0

Nope... LOL

Except in some very specific situations it goes like this.
-------
1. ESR - Original value or less [but not drastically less]
2. Ripple - Original value or more - No such thing as over-kill here.
3. Voltage - Original value - Higher is perfectly acceptable but this in itself isn't an upgrade.
4. uF - Original value. - Fudging uF is discouraged unless there is no way around it.
5. Can size - As long as it physically fits and has the right specs the can size doesn't matter.

Can size discussion is about an easier way to track down replacements that have the other specs you need.
[It's a search aid, not an absolute cross reference.]
WITHIN A SERIES, a larger can will have lower ESR and higher Ripple.

- In looking through data sheets.
If ABC series has an 8x20mm with 1250 mA Ripple
and XYZ series has an an 8x20mm with 1600 mA Ripple
Then XYZ series is a higher grade...
.. and all of XYZ's can sizes will have better specs than the same can sizes in ABC series.
--- so
If you have 5 different can sizes of EFG series to replace and the 8x20mm has 1500 mA -
[reusing the ABC and XYZ numbers above]
Then you know the ABC data sheet is not likely to have any good replacements but the XYZ data sheet is very likely to.
- Thus, you don't need to dig through the ABC data sheet for every cap.

It's not an 'absolute' cross reference though. - There are 'ringers'.
Say one of the old EFG is a 1000uF 8x16mm with 900 mA Ripple.
The XYZ series 1000uF might be 8x11mm with with 840 mA Ripple. [Higher grade but worse specs in THAT uF.]
The ABC series 1000uF might be 8x20mm with 980 mA Ripple. [Lower grade but better specs in THAT uF.]
- These would be cases where can size changes between series over-rule the grades of the series.
- That kind of thing doesn't happen very often but it does happen.
Sometimes it makes life easier and sometimes harder..

A physical inventory is the only way to get it done.
They don't release schematics because they like to keep their trade secrets.

It is critical that you make a cap-map before you remove any caps.
[You don't want to have 3 caps out, get interrupted, and forget which one went where.]
Mark the uF, volts, series, AND POLARITY of any caps you are replacing.
The best way to do this is take a snap-shot of the board and then print it as big as will fit on the paper.
- Then write your notes and mark polarity right on it.
People new to this frequently forget about the polarity. - IS IMPORTANT.
- - - Cap-maps also help in making up your shopping list.

If you MUST change the uF then go 'up' one standard value.
- Caps are +/-20% for uF and one standard value is usually around 20%.
- uF goes down as caps age which means as it ages it will get closer to the original value.

.

[ForumNut]

Re: Ms-7093 ver:1.0

Wow, That was very well explained.

I appreciate very much the time you spent explaining that to me, that's exactly the kind of information I was hoping for.

Perhaps you should sticky that explanation before it gets lost??

I will follow that guide to the letter.

Mapping the caps tonight. I'll let you know how it turns out.

Thanks
FN

[Digital Technophile]

WOW indeed!

Boy I feel so fortunate to be so near to some well-versed folks like PCBONEZ (and others here, too). I'm still trying to wrap my head around this epic disaster in the industry. It seems like almost no one ('cept maybe Rubycon and a couple of others) has been unaffected by this even 10 years later! I mean I bought a Samsung SyncMaster 906BW completely unaware it was full of bad caps (CapXon GL and KM). When it failed, it recapped it with Nichicon HE and PW series caps (long before I joined the forum) I got off of DigiKey. Hopefully Chris P and the crew could get a MediaWiki goin' here, I think that would be an awesome way of documenting the whole of 'lectrolytic caps out there.

Thanks to everyone who contributes and shares their knowledge and experience with us all!

[PCBONEZ]

Re: Ms-7093 ver:1.0

People think of the 'Capacitor Plague' like it's one problem, due to one bug, when in reality it's more like the common cold and there are many different bugs that result in the same symptoms.

One of the biggies that few people talk or know about is that the Chinese mined Aluminum that ends up in CH & TW made caps has different impurities.
Couple that with the fact that the capacitor aluminum made in China are less pure overall and you end up with more & different impurities than are found in Japanese, US, and UK made foils and cans.
Those impurities slowly leach out of the aluminum into the electrolyte ultimately causing the electrolyte to break down.
- The result is the same as the bad caps from the famous 'stolen electrolyte formula' fiasco but it takes a bit longer to happen and if the cap isn't stressed 'worked hard' it might not happen at all due to lower cap internal temperatures.
- THAT problem existed since the dawn of capacitors but it's only in the last 10-15 years that we've 'worked' capacitors hard enough to see large numbers of failures.

The US and UK make very few low ESR caps and they tend to be expensive.
And so now you know why people say to buy Jap caps.
.

[Digital Technophile]

Nice to know

I know when I refer to the plague, I am directly referring to the stolen formula debacle, but without a doubt there has always been some varying degrees due to different manufacturer's formulations of their own "special brew". However I wasn't aware that the specific alloys(?) of aluminum were also a factor. I'm curious as to the type of aluminum that is used as there are many. I have a sneaking suspicion that it is a grade of pure aluminum and not a technical "alloy" like 6061 or 7075, etc. It should be noted that ALL aluminum/alloys form a "skin" of aluminum oxide and is unavoidable barring anodizing. Which further makes me wonder if any of the big boys of caps will ever figure out a clear-coat anodizing with a Teflon seal could prevent these reactions, even with cheap Chinese aluminum; Though it would obviously add quite a bit to the cost of the cap, but might be warranted to specialty and mission critical applications. Definitely a big eye opener, enough in my case that I will be adding an ESR meter to the toolbox when the budget permits (looking at Peak's Atlas ESR atm; any other reputable names would be appreciated..still researching). I know I was surprised to see a few Nichicon series marked as failures (double checked the ones I put in my monitor, to be sure)!

[PCBONEZ]

Re: Ms-7093 ver:1.0

Regarding the Chinese Aluminum you might find this interesting.
CALCE is basically a 'Forensic Failure Analysis' lab at the University of Maryland.
Their whole job is to find out -why- things break.


....

There -are- a few bad Jap caps.

Nichicon HN & HM dated 2001-2004
There was a manufacturing defect in Nichicon HM & HN series made from 2001-2004.
Nichicon admitted so publicly some time in 2004.
Specifically they were over-filled and thermal expansion causes them to bloat or break the seal, leak and then dry-out.
There was also a report later in 2004 that a few runs [not sure how many or of exact dates] of HM and/or HN in which the electrolyte was contaminated with some cleaning chemical.
It's possible some made in late 2004 avoided both problems but since exact dates weren't made public it's assumed all 2004 HM & HN are bad.
So, if the date code is not 2001-2004 HM and HN are fine.
!!! -- 2001-2004 overlaps the 'stolen electrolyte' fiasco and so the two problems are easily confused as being the same thing.

Chemicon KZG & KZJ [and possibly some other less common series]
These are well known to bloat and very often fail completely with no bloating at all.
I suspect it's a chemistry + heat issue with the electrolyte formula used in these specific series. The Advertising [Announcement] that introduced KZJ mentioned it used [not quoting verbatim here] the same advanced electrolyte as their KZG series.
We here at BCN speculated KZJ would experience the same high failure rates as KZG before we ever actually saw any 'in the wild', and we were right.
-
[This is new in the last week.]
Chemicon makes a TMZ series which is a custom/special order cap and not included in their catalog so there are no data sheets are available.
A few days ago I measured the ESR on some TMZ and after adjusting for the can size they fell squarely between KZG and KZJ for ESR grade.
-
Presumably a cap with ESR between KZG and KZJ uses the same 'advanced electrolyte' as KZG and KZJ and thus TMZ would be expected to have the same problems.
-
There is also a 'stealth' [undocumented] series named TMV but as yet no one has directed measured the ESR so the grade is unknown.
The TMV are seen in VRM Vcore so it's -probable- they share the KZG/KZJ type electrolyte.
.
.