Sanyo OSCON Organic Semiconductor Capacitors

[MD Willington]

Since we've been chatting about OSCON's...



Thought a link would be nice..

The application of these has peaked my interest in them...

MD

[willawake]

Re: OSCON link

http://www.semicom.co.uk/manufacture...con_notes.html

woah! oscons.......

i renamed the thread, we will keep this as the main oscon thread

[MD Willington]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Cool..I'm not sure of where to get them other than waifong...

I'm interested in the usage, a lower value used in place of a larger lytic cap...hmm

The values also do not go very high and the voltages are lower, 2.5V, 4V etc...

I have not found any info on the Lelon OCCON's, but the Sanyo's would be my first choice...

MD

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Be very careful about changing capacitance value. ESR is only one parameter, the capacitance is still needed to maintain the designed impedance and storage value. The total impedance(Z) = Xc + ESR (approx) where Xc is determined by the inverse value of the capacitor. Therefore if capacitance is reduced Xc will rise, Xc is in fact the major determinant of impedance. ESR becomes important at high ripple currents where loss and heat are generated due to the ESR value.

Comparing ESR values the Rubycon MCZ series often has a lower value than OSCON with lower cost and slightly better availability.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Here's a US distributor that has Sanyo Oscon's, as well as regular Sanyo electrolytics..

http://www.tenfourltd.com/oscon/oscon_spe.html


I've ordered some Sanken Power transistors from them for my work, and the parts were new and in factory packaging.

[willawake]

Re: Sanyo OSCON Organic Semiconductor Capacitors

good find. what is their minimum order value though? will they deal with individuals?

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

good find. what is their minimum order value though? will they deal with individuals?

Not sure. I think we ordered 100 Sanken power transistors the last time. I guess the best way to find out would just be to send them an email. They have a contact form.. The actual website is tenfourltd.com , i just linked to the SP series OSCON's, as i'm guessing they're the most relevent OSCON for a motherboard.

[Per Hansson]

Re: Sanyo OSCON Organic Semiconductor Capacitors

It would be interesting if someone tested using OS-Con's on their mainboard instead of normal caps to see if it increases stability/max stable overclock speed...

Something tells me it might, their specs looks impressive!

[willawake]

Re: Sanyo OSCON Organic Semiconductor Capacitors

i believe that replacing all the caps with a good brand will improve overclocking stability. you can see some of the review boards favour oscon sepc for vcore and i think that is the way to go. some boards use a mix of rubycon mcz, the yellow fujitsu fp-cap, nichicon HM HN etc for vcore. sepc would be superior but it really depends on the cost/performance, how cheap you can obtain them if at all.

although there are more things to do, get a good psu and attend to the caps in that, then look into superior northbridge cooling, heatsinks for the mosfets, superior cpu cooling, case cooling. in the end though it will be limited to what your cpu, ram, board can do and of course water cooling will always be superior to air.

i am not really interested in overclocking though, i may do some tests to see what is possible with regard to caps when i recap this albatron board.

you can see some possibilites here but it is more advanced.

http://www.finetune.jp/~lyuka/interests/pc/p4c800e.html
http://www.finetune.jp/~lyuka/interests/pc/bp6.html

i have been reading documents about power delivery to the mobos and although i am not very proficient in electronics i can see that care must be taken in selection of caps and their values for best performance although if stability and longevity is the only concern then good caps in reasonably similar values are a fine choice.

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

In capacitive filters the size of capacitance determines the ripple voltage and thus the ripple current. OSCON caps have great ripple current ratings but the impedance is about the same as Rubicon MCZ. This means that the OSCON caps technology can withstand the higher dissipation of the greater ripple value. If OSCON is substituted for MCZ using the same value there would be no gain, the extra ripple rating of OSCON would not be utilised. If lower value OSCON caps are used then the ripple voltage will rise and the OSCON caps will hopefully be within rating, however is it worth raising the designed ripple voltage level and reducing surge capability just to utilise an expensive device? OSCON caps offer opportunity for designers to do the job properly.

When the new SAMXON GA series capacitors are available and tested in the field they would seem to be more cost effective. Their ripple rating is high with an ultra low impedance, better than OSCON. Take a look at the SAMXON series at:
http://www.samxon.com/Product.php?La...ap=3&LoadLife=

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Take a look at the figures of comparison OSCON to others
Product Capacitance Z (100Khz) Ripple rating ESR mOhm

Rubicon MCZ 2200uF 6.3 --11.02--- 2770 (105C)----- 11
OSCON SP 2200uF 4.0 ----10.03---- 7100 (45C)------ 10
SAMXON GA 2200uF 6.3-- 10-------- 2960 (105C)------ 9.97

Well try as I may the figures above bunchup so have inserted - as seperator


Note the OSCON ripple rating is at only 45C. At 105C this falls to 25% ie 1775mA !!!! OSCON endurance is no longer than the others. So apples for apples they are not that good.

The pure capacitor Xc for 2200uF at 100Khz is 0.723 mohm seeing the ESR is typically 10 mohm it is the major component i.e Xc affect the total impedance (Z) at 100Khz by only 0.3% (using the correct formula, not adding).

To put my comments and data simply. It is not worth replacing caps with OSCON. They are best applied to new designs that apply them correctly. And are they really any better????

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Now for the even bigger surprise from SAMXON.
I quoted the GA series 2200uf 6.3 10mm x 16mm (small) 10mohm and 2960 mA

NOW the GA 2200uF 10mm x 20mm is an amazing 7mohm and 3770mA at 105C

Hard to beat. Input please.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I still think the Oscon's are about the best kind of cap you're going to put on a motherboard. For one thing, if you're running your caps at 105 C, none of them are going to be particularly happy.

Given that Electrolytics tend to dry out, and that their life is derated by such high temps, the fact that Oscon's are polymer inside, so it would seem to me that the Oscon's would be more stable in extreme conditions, or even in not so extreme ones. I've never cut an Oscon open, as i've never met a failed one in service. There's a reason that high end boards have some polymer caps on them, and long term stability is one of them.

Oscon's are proven good technology caps. I'm not even sure how much the place charges for them. I recovered a set of 330uF Oscon's from a 1996 vintage Intel VS440FX Pentium Pro motherboard about two months ago, that had seen a lot of use, and Pentium Pro's weren't known for being cool running chips. The whole reason the board failed was that i was running a PPro/200 with 1M L2 at 233Mhz (over 60W of heat there), and the VRM was only designed for a 200/256K cache model, roughly 35W of power. The switching FET on the board literally split in half from the sheer amount of heat.

After almost 10 years, these caps are still well within tolerance, and i'm sure they're just as good in other ways. This goes for quality Electrolytics also, such as Nichicon UPW. I've got a '97 vintage AL440LX with Nichicon UPW, and they're still in quite good shape also. Hence why i use the large diameter UPW's for recapping, if they'll fit. Longer life (>5000 hours) and known reliability.. I'm sure that other Rubycon, Sanyo, and Nichicons should have at least half of the runtime, in reduced diameter packaging anyway, and still work well.


I'm sorry, but Samxcon is not proven to me, and Intel doesn't put them on their motherboards, so i doubt i'll be recapping boards with them. The cost of the caps isn't much compared to the cost of the time involved in doing a good job of recapping a board, so i use the best i can.


As for filtering ripple, the big caps don't do that. The frequency of the switching FET's is much higher than what kind of frequency a 1500 to 3300uF cap would filter. If we were talking about filtering the output of a half or full wave bridge, that would be a different story, as that would either be 60 or 120Hz, not the 20kHz + that we see in the switching FET's in the mobo VRM.

[gg1978]

Re: Sanyo OSCON Organic Semiconductor Capacitors

Here's a bit more proof that Oscon's are the best. Here's a link to an Onsemi VRM 9.x controller that has a reference design in it, as to how the make the VRM work.

https://cdn.badcaps-static.com/pdfs/...d216e16b7d.pdf

It's a two phase VRM controller.

Here's the options for the output caps, as listed in the PDF.

1. (10) Rubycon MBZ 1500uF @ 16VDC total of 15000uF output capacitance.

2. (8) Sanyo Oscon 820uF @ 4VDC, total of 6560uF output capacitance.

As for why they chose the 16V variety of Rubycon MBZ for this ref design, i have no idea. A set of 6.3V ones would have worked fine, and would probably be smaller in diameter, unless it was to match the diameter of the Oscon's..

In any event, in a reference design, you don't want to have your stuff look bad, so it has to meet Intels VRM spec to be "good".


The Oscon's are better, as they're obviously using the extra output current capability of the Oscon's to get away with using 2 less caps, and cut the capacitance of each output cap from 1500uF to 820uF.

I've seen this trend in other VRM controller datasheets, so it's not unique to ON Semiconductor.

[davmax]

Re: Sanyo OSCON Organic Semiconductor Capacitors

First of all I agree that Samxon are as yet not proven in the field, but I think Rubicon MCZ are.

I have compared capacitors with the same ESR value, therefore apples for apples comparison.
Now I am sticking strictly to the manufacturers specifications, no assumptions. The figures I extracted a posted are just that. They show that the main OSCON advantage is that at lower temps higher ripple currents can be sustained. If you can keep temp down to 45C they are really excellent, at 85C they fall of to 40% rating ie the 7100mA value is down to 2840mA. This is comparable with the Rubicon and Samxon values at 105C.

Now as for life time, again from specs. The OSCON life is stated as 2000hrs at rated voltage and 105C with no mention of ripple current value. Both Rubicon and Samxon state 2000hrs at rated voltage at 105C with max ripple current.

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.

The design example quoted Rubicon MBZ. These are 40% worse than MCZ and this could account for the difference in OSCON to Rubicon cap count.

In summary OSCON have had a good name for a long time but other manufacturers have responded with improved components. Oscon still leads for low temp ripple but I can see no other specification that justifies their use.

There is a need to remain factual and what better than the spec claims of the manufacturers.

I am keen to test Samxon the manufacturer is long standing and focused on quality. Big Pope has been supplying these to over clockers with very good reported results.

If there is a fact that I have missed I would be happy to learn.

[Big Pope]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I agree with davmax.

I don't suggest to use OSCON in motherboard. Actaully, it is meaningless compare with Rubycon MCZ, Nippin-Chemicon KZJ, Samxon GC and Nichicon HN. It is because OSCON is really very expensive, but without much improve for the performance of motherboard.

Actaully, many of high end display cards using with OSCON. It is because display card for high end PC game users. They are willing to pay with expensive price. That's explain one of reason why OSCON will use in display card, but a little factory will use OSCON in motherboard.

[Big Pope]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I still think the Oscon's are about the best kind of cap you're going to put on a motherboard. For one thing, if you're running your caps at 105 C, none of them are going to be particularly happy.

Given that Electrolytics tend to dry out, and that their life is derated by such high temps, the fact that Oscon's are polymer inside, so it would seem to me that the Oscon's would be more stable in extreme conditions, or even in not so extreme ones. I've never cut an Oscon open, as i've never met a failed one in service. There's a reason that high end boards have some polymer caps on them, and long term stability is one of them.

Oscon's are proven good technology caps. I'm not even sure how much the place charges for them. I recovered a set of 330uF Oscon's from a 1996 vintage Intel VS440FX Pentium Pro motherboard about two months ago, that had seen a lot of use, and Pentium Pro's weren't known for being cool running chips. The whole reason the board failed was that i was running a PPro/200 with 1M L2 at 233Mhz (over 60W of heat there), and the VRM was only designed for a 200/256K cache model, roughly 35W of power. The switching FET on the board literally split in half from the sheer amount of heat.

After almost 10 years, these caps are still well within tolerance, and i'm sure they're just as good in other ways. This goes for quality Electrolytics also, such as Nichicon UPW. I've got a '97 vintage AL440LX with Nichicon UPW, and they're still in quite good shape also. Hence why i use the large diameter UPW's for recapping, if they'll fit. Longer life (>5000 hours) and known reliability.. I'm sure that other Rubycon, Sanyo, and Nichicons should have at least half of the runtime, in reduced diameter packaging anyway, and still work well.


I'm sorry, but Samxcon is not proven to me, and Intel doesn't put them on their motherboards, so i doubt i'll be recapping boards with them. The cost of the caps isn't much compared to the cost of the time involved in doing a good job of recapping a board, so i use the best i can.


As for filtering ripple, the big caps don't do that. The frequency of the switching FET's is much higher than what kind of frequency a 1500 to 3300uF cap would filter. If we were talking about filtering the output of a half or full wave bridge, that would be a different story, as that would either be 60 or 120Hz, not the 20kHz + that we see in the switching FET's in the mobo VRM.

Do you want to test SAMXON GD and GC capacitors? I can send you some samples free of charge. Please PM me if you want. Thank you.

[Rainbow]

Re: Sanyo OSCON Organic Semiconductor Capacitors

I've seen OSCONs even on some not-so-high-end boards too - but only one per board . One on LuckyStar Socket A board and one on ECS/PC Chips Slot A board.

[kc8adu]

Re: Sanyo OSCON Organic Semiconductor Capacitors

i have seen oscons fail.i posted the pics .
they were ucc if i recall.and i save oscons from junk boards before i pitch the carcass.never seen a sanyo out of spec.got some 1000@16 out of a proliant mainboard yesterday.these are huge!.it was something other than space saving they were after.

[willawake]

Re: Sanyo OSCON Organic Semiconductor Capacitors

it was something other than space saving they were after

i have seen a few nocona chipset servers HP and Dell using exclusively oscon SP.