Recap; thinking of using Sanyo OS-CON

[Per Hansson]

Hi, I bought a Epox 8RDA+ for 5$ knowing it's cap's where bad...

I've located replacement caps on elfa.se. They are Sanyo MV-WX types...

However I was thinking of using Sanyo OS-CON caps instead, just for the heck of it ;-)

But I need to know how many farads they should be... The caps on the board that I'm going to swap out are:

2200uF 10v 4x (25x10mm)
1500uF 6,3v 5x (1x 15x10, rest 21x10mm)
1000uF 6,3v 9x (15x8mm)

I found an old thread here on the forums where a GFX card manufacturer had chosen to use 360uF OS-CON caps in the place of normal 1000uF electrolytic caps (without modifying the rest of the cards components or PCB in any way)

So does someone have some chart or description on how to interchange these cap types?

[AK0R]

Re: Recap; thinking of using Sanyo OS-CON

There is no such chart/description. Capacitance is capacitance, regardless of the type of capacitor being used. Different types of capacitors have different capabilities (working voltage, ESR, slew rate=charge/discharge rate, etc.), but in the end, the circuit design relies on having a particular amount of capacitance. In filtering applications, which is the case for most of these capacitors, more is usually better. It's a stretch to say that "no other changes were made" without the schematics of the two versions to compare, but assume that's the case. Suppose the design requires at least 300uF of filtering. Because of economies of scale in purchasing, the manufacturer originally elects to use 1000uF aluminum oxide electrolytics, of which he has a ton but are hand-loaded, rather than buy a new value (330, 470, 560, or 820uF) that he doesn't currently stock for any product and would also be hand-loaded. Later, a new product (surface-mount cans or OSCONs) becomes available that can be machine-loaded. Now, since a new product is being purchased, a value closer to the requirement is more likely to be cheaper, so the value "changes, with no other changes to the card". In both cases, the installed part meets the design requirements. But which one is better? If the parts are equally reliable, the earlier version provides more capacitance, and may actually be more stable. The point is, without having the schematic and knowing the design requirements, no such judgment can be made, and the only reasonable course of action is to replace bad parts with new ones having equal or better ratings.

[Per Hansson]

Re: Recap; thinking of using Sanyo OS-CON

Thank you AK0R for that explanation

I now better understand how this relates... Also saw this page: http://www.diyparadise.com/oscon.html

According to that page a OS-CON at 47uF does an as good job as a regular cap rated at 100uF does...

But I'll just get some regular replacement caps of good brand (Sanyo MV-WX) but now I atleast feel like I know a bit more about this ;-)

[Per Hansson]

Re: Recap; thinking of using Sanyo OS-CON

Recapped the board yesterday, it's working perfectly fine now...

Voltage on middle leg of CPU VRM's are 1,71v and 4,86v respectivley, on the filter caps it is 4,82-4,86v this is running 3DMark01 so stress on CPU is high... Since the voltage on the caps are fluctuating slightly but not the voltage on the VRM's I'm assuming the caps are doing their job now... I expect that had I done this measurement with the bad caps the voltage would have fluctuated on the VRM's?

The coils and VRM's are below 60°C after hours of 3DMark looping, Memtest86+ passed 12 hours of tests...

I'm running the CPU at 200x10,5 which should really stress this poor board with no fan on the northbridge...

All in all I think this was a very successfull repair work.

[willawake]

Re: Recap; thinking of using Sanyo OS-CON

that row of teapo gonna get replaced also?

i have heard the same for oscon, about 50% repacement uf value and same esr but never seen any application notes in that respect.

[kc8adu]

Re: Recap; thinking of using Sanyo OS-CON

i had some 820@4v oscons i put in a supermicro 370dl3 near the cpu sockets. that is on vcore
vcore was dead clean with a scope when i was done.
i replaced 1000@10v

[kc8adu]

Re: Recap; thinking of using Sanyo OS-CON

that row of teapo gonna get replaced also?

you can change em now or wish you had later.
teapo's can look good but be open.thats >20 ohms esr.
had some ibm thin clients come in for lockup,no post,ect that were full of teapo's
a few pokes with a scope probe told me they had to go.

[Per Hansson]

Re: Recap; thinking of using Sanyo OS-CON

Regarding the TEAPO caps: I had no idea the caps could look fine but be dead...

Unfortunantley I do not have a scope nor the knowledge to know what cheap scope would fit my needs of eBay... So I'm stuck with my multimeter...

But I mean the GSC caps looked really crap, and these TEAPO caps looks perfectly fine... Plus my now almost 2 years old highly OC'ed and quiet (read very hot) Epox 8RDA3+ uses TEAPO caps for everything and they look fine to me...

[Per Hansson]

Re: Recap; thinking of using Sanyo OS-CON

Measured the voltage on the TEAPO caps and they where 1,71 volt not fluctuating, just like the VRM's...

BTW, at what microfarad rating can I expect caps to not bulge anymore, someone told me in another post 470uF and below, the black cap you see in the picture at the top is 470uF 16v TEAPO...

[AK0R]

Re: Recap; thinking of using Sanyo OS-CON

BTW, at what microfarad rating can I expect caps to not bulge anymore, someone told me in another post 470uF and below, the black cap you see in the picture at the top is 470uF 16v TEAPO...

470μF and above should be replaced for sure. Some users have reported problems with caps as low as 220μF. It depends on the cap manufacturer on whether the lower values (220 to <1000) are any good, and no, I can't tell you which ones are good and which aren't.