Sanyo OSCON Organic Semiconductor Capacitors

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Sorry, wrong:

quote You are very incorrect in saying that the large capacitors are not filtering 100 Khz. The very reason that they must have low ESR at that frequency is that it is very necessary. quote

Check out page 17 of the VRM controller datasheet i linked.:

Under design Procedure,

The "bulk" capacitors provide "hold up" during transient loading. The low impedace, high frequency ceramic capacitors reduce steady state ripple and bypass the bulk capacitance when the output current changes very quickly.


True, the caps do "filter" the output in a way, but not in the way you say.



The low ESR of the bulk caps is to reduce heating in them while they're at work, not for bypassing the 100kHz + switching noise. That's what the ceramic caps do. The lower the ESR, the less the cap will heat up, given a certain amount of current.


I'm not saying here that the Sanyo Oscon's are the best deal for your money. But i'm saying they're about the best cap you can put on a motherboard's switching VRM. They can supply more current than most other caps, provided you keep the temps down. If i was overclocking a liquid nitrogen cooled Pentium 4 Prescott at say 4.5Ghz or higher, i would certainly want Oscon's in my VRM, since such a Prescott would be extremely demanding on a VRM, and any voltage droop would be a bad thing..




Kc8adu:

Were the failed Oscon's Sanyo, or do other people make Oscon's also?? I thought that only Sanyo made the ones with the Oscon brand name.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Here's the only modern board that i've found that ONLY uses Polymer caps on it. Of course it's a ~$630USD dual Xeon motherboard, but it shows that Oscon's do have their place.

http://images10.newegg.com/NeweggIma...121-218-01.JPG

The purple and blue ones are Oscon's, while the yellow ones are Fujitsu polymer's..

[kc8adu]

Re: Sanyo OSCON Organic Semiconductor Capacitors

https://www.badcaps.net/forum/showth...ighlight=oscon

[Chris1992]

Re: Sanyo OSCON Organic Semiconductor Capacitors

it's a 1u board, good call on their part b/c 1u is very hot form factor.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

kc8adu:

I'd say that looks like a United Chemicon Logo. The Oscon's i salvaged from the '96 vintage VS440FX are a rather purple color, labeled "Sanyo OS-CON 330/6.3 105 C 8A ..


I also found an EVGA board (their only one) at newegg that looks like it has a set of OSCON's on the output of the CPU VRM, although i can't be sure from the picture.
Could be wrong on that, and they're just low profile regular electrolytics. But since there's plenty of room for regular profile electrolytics on that board, that's why i think they're Oscon's. They also don't appear to have the vent stamped into the top of the cap. Regular Sanyo electrolytics have the very distinctive "K" vent on them. Rubycon has a different "K" vent, while United Chemicon has a 3 pointed vent, and Nichicon's usually have a 4 pointed vent. IIRC anyway..

http://images10.newegg.com/NeweggIma...188-002-02.JPG

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I joined this thread because there was a clear preference for OSCON caps. Was it because they are a very good product? Was it because it was safe to use a quality brand having been caught with bad caps i.e. good insurance? How did alternates stack up and were they sufficient, economic and reliable?

Looking at the 2200uF level comparison there appeared to be not much in it. However OSCON technology has some very good characteristics no better defined than in the SEPC series. They have high ripple current spec'd at 105C, low ESR at a lower capacitance than aluminium caps and they are small in dimensions. The price at Wai Fong is 330% higher than a Rubicon MCZ of equivalent ESR performance but with 40% less ripple rating.

gg1978 kindly put forward a reference design. https://cdn.badcaps-static.com/pdfs/...d216e16b7d.pdf last updated in 2003. The recommended output capacitors were;
8 only Oscon 820uF 4V or 10 only Rubicon MBZ 1500uF 16V.

The 16Volt MBZ was chosen so that the ESR is the right value (MBZ is superseded by MCZ).
Analysis of this design does not include ESL because it cannot be determined. In determining ripple the impedance Z of the bulk caps and the ceramics are seen as a parallel to the load and sharing the ripple in proportion to their values. In the case of the bulk caps the parallel ERS values combine to produce a value between 1.3 and 1.5 mohm taking a bulk of any 220kHz ripple the ceramics pickup the harmonic energy. Below are the differences between these caps.

OSCON performance. Bypass 78% of ripple at 220Khz. Each cap has a 25% higher ripple than the MBZ caps. Any 120Hz ripple present would be about 200% higher in the load due to the reduced capacitance. Because there are eight caps there is also a 17% less transient current carrying ability (ref. the formula page 17 of reference design).
MCZ performance. Bypass 80% of ripple at 220Khz.

The ceramic capacitors are passing about 19% of ripple at 220Khz this added to the figures above indicate about 1% ripple in the load. The ceramics are intended to bypass the even higher harmonic ripple currents. All capacitors respond to surge demand.

The above was a 2003 scenario and both OSCON, Rubicon, Nichicon and Samxon have advanced since that time.

So today the caps in that same design could be:

Rubicon MCZ quantity 8 of any of the following:
1. 1800u 6.3V
2. 1500u 12V
3. 1000u 16V

OSCON SEPC quantity 5 of 820u 4V

Samxon GA quantity 8 of any of the following:
1. 1000u 6.3V
2. 820u 12V

Or possibly 5 only 2200u 6.3V or 10V depending on absolute ripple values (max rating is 40% less than SEPC)

A point to reinforce is that at these high switching power supply frequencies the prime parameter is ESR as the capacitive reactance is negligible. Therefore the value of capacitance is not critical unless lower frequency surges or 120Hz have to be accommodated.
It should be noted that the high ripple rating of OSCON it not always good to utilise because if the ESR is unchanged a high ripple value in the cap will also mean higher ripple voltage applied to the load. Ripple rating is based on thermal performance.

Clearly investigation shows that the OSCON capacitor to still be superior at a premium price. They are even specified up to 500Khz where inductive reactance really starts to kick in. For new designs operating up to 500Khz I would expect the multi layer ceramics to slot in here. They would be small and cost effective. When using OSCON to replace aluminium caps care should be taken when reducing the capacitance value if the design criteria are not known or any 120Hz component is present.

No matter what technology is used the requirement is that the performance and reliability are sufficient. It can be over kill to replace aluminium caps with OSCON if the equipment operates without problem, there may be no performance gained. It is certain that manufactures will continue to look at cost.

Rubicon MCZ, Nichicon HZ, and Samxom GA have sufficient specification for most motherboards today. The simple method is to replace spec for spec.

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Thanks for your response gg1978. Page 17 of the reference design may or may not be misleading. Certainly energy storage is required for transients but there is ripple passing through the bulk capacitors the rms value of which causes the heating in the ESR (this is the working heat, capacitance itself does not generate heat). Typically steady state ripple causes the heating. Now the amount of ripple depends at what frequency the major ripple power exists. For example it may be the case that this power resides at the 3rd harmonic (660Khz) that being the case the impedance of the bulk capacitors will rise and that of the ceramics will fall thus shifting the ripple current more into the ceramics. Transient energy can certainly form part of the heat in the ESR, but usually a minor player.

It occurs to me that if the major ripple is in fact being diverted to the ceramics the ripple rating of the bulk capacitors is not very relevant and the low ripple rated aluminium caps will have no problem.

As stated in the reference design there are so many variables, some unknown, that it is not possible to calculate exactly what is going on. So during product design simulation or better a prototype based on approximations is measured and the design then tuned.

For comparison of capacitor performance the task is made simpler by picking one frequency and a fixed load. The load is best the rated load where ripple is highest. In the example 1.6V at 45A is represented by a resistive load of 35 mohms. Under these conditions a good comparison can be made. But it is only representative of one frequency of 220Khz. I actually calculated at 120Hz as well for comparison.

Hope this is helpful.

[MD Willington]

Re: Sanyo OSCON Organic Semiconductor Capacitors

EVGA/Jetway main board, I know the Jetway does not use Sanyo's, so I doubt the EVGA board would use them too... but maybe they do, I don't have one of those boards to scrutinize.

Jetway A210GDMS-Pro
http://www.lelon.com.tw/english/product-e/occon-e.htm

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Concluding my close look at the application of OSCONS compared to aluminium caps.

OSCON are undoubtedly a top product but to obtain performance advantage they need to be applied correctly.

Take the example of the reference design linked by gg1978.
The OSCON gain in component count thus removing part of the extra cost. But in terms of performance there is no gain, in fact using the formula supplied there is a 17% loss in transient performance. The formula only accounts for ESR it does not account for loss of bulk capacitance, in this case a 56% loss or conversely the Rubicons have 229% more bulk capacitance. This means that there will be further deterioration in terms of energy loss i.e. voltage drop. To emphasise this point consider the ESR without any bulk capacitance, so as capacitance decreases both storage voltage drop and ripple increase. In summary this particular design is probably sufficient but it can be seen that the OSCON caps do not improve performance in this case.

So OSCON caps need to be applied correctly, reducing the component count or the capacitance is not helping to improve performance. Make sure this is taken into account.

A general rule would be keep ESR low without too much reduction in capacitance.

It would seem from the evaluation that the high ripple rating may not count for much, this is design dependent. To minimise ESR multiple caps are used, they share ripple and there may be ceramics sharing as well. Pushing the ripple current up to take benefit of the higher OSCON ratings will mean only one thing, high ripple voltage and more ripple in the load.

In conclusion it is certainly fine to use high quality aluminium caps that have an equal ESR (usually with more bulk capacitance). Manufacturers have been using higher ESR caps in larger numbers to achieve the same result but at lower cost.

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

See this link: https://www.badcaps.net/forum/showthread.php?t=1305

[larrymoencurly]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I saw some OSCONs in the CPU voltage regulator of a lowly Gateway 466 MHz Celeron system with integrated graphics. The other caps were Chemicons.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Yeah, it's probably an Intel OEM motherboard though.. I've seen them on SE440BX, SE440BX2, and SE440BX3, as well as the VS440FX and some of the old slot one Gateway "Jabil" slot 1 boards. Funny, i've seen a good mix on intel boards, from Rubycon to Nichicon UPW and chemicon, and of course the Oscon's on Vcore, right by the CPU slot. Just got a SE440BX with Panasonic's though, first time i've ever seen an intel OEM board with Pannies on it..

I saw some OSCONs in the CPU voltage regulator of a lowly Gateway 466 MHz Celeron system with integrated graphics. The other caps were Chemicons.

[willawake]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Oscons on the cpu board of an HP Netserver 5 100/LC. 100Mhz Pentium. Jeez Oscons are older than i thought.



lotsa tantalum also



wah!...monster folder...well makes a good back scratcher anyway



oscon together with panny HFQ

[Rainbow]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Yes, OSCONs aren't new. They were used in notebooks and servers for at least 10 years. I have some in 1997 notebook.