The Great Capacitor Showdown - Rubycon vs Panasonic vs Samxon

[Topcat]

The much awaited showdown is being prepared. At long last, we will see how Rubycon, Samxon, and panasonic will fare against eachother in a head to head comparison over an extended period of time.

The scenario:

3 identical systems (Abit KT7-RAID motherboards, all running the A9 BIOS), one running Rubycon MBZ series caps, one running Samxon GC and GD series caps, and the third running Panasonic FC series caps. All three system will have the same CPU's, amount and brand of RAM, PSU's, Operating Systems (WinXP Pro) and cases. When the test is underway, these systems will be exposed to inadequite cooling (not enough to cause overheating, but enough to stress things a bit). They will be running F@H applications for the Badcaps.net Folding Team, with the app set at the highest priority to keep them busy and working hard. Periodically, I will shut them down and take ESR readings from all caps on each board, as well as monitoring temps and stability. The first one to drop out of ESR specs, or fail completely will be eliminated, and the brand scourned and laughed at here.... I plan to run this test for atleast a year, reporting on it every couple weeks to record deterioration and changes. Honestly, this should be a good comparison, and I really don't forsee any of the brands swelling or venting their innards. This will be the 'test of time' comparison, and may the best brand win. All the brands will deteriorate a little over the course of this, the one that does it the least will immerge as the winner!

I purchased the Rubies directly from Rubycon. The Pannies will be coming from Digikey (I still have to order the right values for the KT7). The Samxon caps were donated by Joe (Big Pope). I would like to thank Joe for this contribution!

Here's some pics of the KT7 recapped in Samxon caps as it s recapping was completed, and I did the initial testing with it. The board passed initial testing with flying colors.

I will get some pics of the board capped in Rubycon MBZ's tomorrow. I'll also locate and order the Panasonic FC's as well.



























More updates soon........

[PeteS in CA]

Some random suggestions and comments:

* Please try to track location wrt hot components in your data; there may be interesting differences within series that are related to the parts' immediate ambient;

* I suggest using a spreadsheet, recording start date/time, measurement date/time, calculated cumulative hours, and ESR; be sure to take pre-test data;

* Unless a DUT vents first, you might want to consider doubling the initial ESR a technical failure;

* C and ESR are very temperature sensitive; take your ESR readings after allowing 1-2 hours of cooling (the point is to cool the core, not just the can); recording case temps might be useful, but possibly TMI for your purposes.

This test sounds really good.

[gonzo0815]

I only can say awsome TC, this is absolut an interesting trial. I am very excited wich brand will win and if those Samxon can stand up with those Rubycon`s. Very aprecitated and TX for the effort making us a little bit smarter.

[Tom41]

Just out of interest, what were the original caps on the new motherboards before you recapped them? It might also be a good idea to include Nichicon caps in a similar test, as they are renowned for being 'good' caps. The spate of failures was limited to a certain batch of caps, and the problem should not come up again.

Next, do a battle of the bad caps (such as Tayeh and GSC) - see which ones vent first!

[gonzo0815]

Originally Posted by Tom41

Just out of interest, what were the original caps on the new motherboards before you recapped them? It might also be a good idea to include Nichicon caps in a similar test, as they are renowned for being 'good' caps. The spate of failures was limited to a certain batch of caps, and the problem should not come up again.

Well, i think Nichicon isn`t bad aktuall, but some batches are certainely bad. Up to now, i don`t have a bad feeling about those Nichicon HM(N) caps in my new Gigabyte GBU8x board.

Originally Posted by Tom41

Next, do a battle of the bad caps (such as Tayeh and GSC) - see which ones vent first!

Who is interested in crap vs crap?
Both brands your mentioned are absolut BS, why bother with those caps?
No seriouse recapper will use them for any purpose.

The only other brands wich will be intersting could be Capxon and may be Jamicon, but in any way, i have enough faith in those to use them without such an test.

[Shroomie]

...Maybe Teapo as well - they seem to have improved.

Tayeh vs. GSC would be fun just to see which truly is the worst of the worst. =D

[Rainbow]

And what about Chhsi?

[tiresias]

Originally Posted by Rainbow

And what about Chhsi?

That would be a funny test - I think best contenders for the "worst" prize would be Chhsi indeed (these seem to fail almost all the time), Rulycon (how dodgy can you get?), the TMS that Topcat posted, and the mysterious "SG" that our user "pinobot" found were starting to vent even before the machine was actually used.

Fuhjyyu is probably not the very worst of the worst (heck, at least they have a website), but it has probably caused more grief than most due to such widespread use in ChannelWell PSUs...

[yanz]

It sounds fun I can't download those pics now as i still in the area where only dialup available, can't wait till i get home and see those pics!

Someone mentions about Teapo? I told my friend to buy me some teapo caps but somehow they are dissapeared from the market.. I want to install them on some asus p4 board with ost caps. That would be Teapo vs Panasonic FJ (identical boards).

BTW, on my KT7 i have installed four Samxon GC caps (in the vrm input). The rest are Sanyo WG and some other jap caps. The PSU is FSP300-60BT with Nichicon HE and Sanyo WG in the ouput filter. That KT7 is a tough board.

[nsx29]

Any Updates?

[tazwegion]

Originally Posted by Tom41

Just out of interest, what were the original caps on the new motherboards before you recapped them?

From memory, both my Abit KT7's were covered in Jackons (grey/silver & green jacket)

This is an excellent idea TC... the term 'showdown' conjurs up images of gunslingers in the wildwest!

[nsx29]

There is something that i do not understand about samxon caps. Looking at a lot spec. pdf, japanese cap manufacturer (nichicon, chemicon, panasonic) don't seem to make 8mm diameter 1.5cm tall 1000uf 6.3V caps only 820 uf.

So I wonder if samxons are over rated in some way.

[davmax]

A possibility that Samxon may be slightly over rated . Not huge difference from 820 to 1000.

All samples working well to date. Testing continues.

[Spacedye69]

Rubycon mbz 1000uf 6.3v 8mm x 11.5mm.

[trodas]

Suggestion - use entierly same bios settings and same memory modules too - same size, same bank and same settings to ensure the comparable operation

...and hey, Samxons got a disadvantage already! They was running already, while there Rubyes and Nichicons resting in stores yet

And one more question. The color of the IDE ports are somewhat different on the right mobo - why is that? Is the HW versions of these mobos identical?

And the missing Vcore final cap are part of original design??? That suxx anyway...

------------------

Shroomie -

Maybe Teapo as well - they seem to have improved.

Don't make me laugh. Teapos was used into the pre-testing boards A9RX480 Grouper by Sappire and failed miserably within few weeks for many reviewers around the world! Improved? Blah!


Rainbow -

And what about Chhsi?

Now your reply make me laugh loud You can't be serious - Topcat sure ain't going to find a source for such inferior crap...


nsx29 - I doubt it, since there actually are even smaller 1000uF caps, not to mention that for example Pannyes are usually bigger caps that all others. A nice example - Panny FM 470uF 16V is d8x16 while Samxon GC 470uF 16V is d8x12.
Both d8 are more likely 8.2mm in reality with the PVC enclosure anyway.

So, in the size Samxons GC = Rubycons MBZ. Specs favorizing Samxons, tough. By a huge margin:
Samxon GC 1800uF 16V - impedance 0,009 - ripple current 3190 (2 000 h) 10x25 - (2200uF 0,008 3370mA - 12,5x25)
Rubycon MBZ 1800uF 16V - impedance 0,013 - ripple current 2800 (2 000 h) 10x23

Ruby MCZ compete well with GC:
Rubycon MCZ 1800uF 16V - impedance 0,009 - ripple current 3230 (2 000 h) 10x25

However are nowhere near to be found except few pieces from WaiFong.hk ...

[Big Pope]

Originally Posted by trodas

And one more question. The color of the IDE ports are somewhat different on the right mobo - why is that? Is the HW versions of these mobos identical?

And the missing Vcore final cap are part of original design??? That suxx anyway...

The colored IDE ports are HighPoint 370 RAID.

[trodas]

Well, yes. Sure they are RAID ports, and they are (with the chipset that control them, clearly marked...) on all mobos. However my question was why these colors aren't identical (as they usualy are) or and if these mobos are the same HW revision of the PCB
The bottom line of course is, that different mobo revision could have some impact on the test. Probably low, but... who can say that for sure? We don't even know if the impact would be good or bad...

Sure, with limited resources and so on, we probably can't get ideal testing, however we should try our best, should we?

After all, we should stuff this one with Nichicons, where we can afford some uncertainity, and not Rubycons to get at least the comparsion with Samxons and Rubycons straight as much, as we can

[Topcat]

I was given a 4th KT7 this morning. Is there one more brands we'd like to add to this experiment? Sanyo (my next favorite to Rubycon)? Nichicon maybe? I'm not a big fan of Nichicon after the HN and HM mess.... opinions?

[Spacedye69]

I would like to see Nichicon to clear up the HM HN mess.

[Topcat]

I'm actually leaning more toward Sanyo WG's over Nichicon's. I don't have a source for them though....