4700uF 6.3v and 3300uF 16v in a 10mm Diameter

[mockingbird]

Originally Posted by Topcat

I thought RS's were discontinued....I haven't been able to get them since Joe Lo (bigpope) went MIA. I do like the specs of the KZN better for sure though.

I'm still in touch with Joe on occasion. You can still reach him through the Chinese apps. Last I spoke with him he said he could still get Samxon but I think I decided to go with local alternatives at the time instead.

Yea Behemot's RS are legit and newer stock than Joe's. He did a custom order on them as well a while back. I used them in a Socket370 VRM high last week along with some NOS Nichicon HZ for the VRM low.

[RJARRRPCGP]

Originally Posted by mockingbird

I'm still in touch with Joe on occasion. You can still reach him through the Chinese apps. Last I spoke with him he said he could still get Samxon but I think I decided to go with local alternatives at the time instead.

Yea Behemot's RS are legit and newer stock than Joe's. He did a custom order on them as well a while back. I used them in a Socket370 VRM high last week along with some NOS Nichicon HZ for the VRM low.

NOS?=Of course NOS, FTW!

[momaka]

Originally Posted by Per Hansson

Yea I would also not use KMG in a modern PSU, General Purpose caps do not belong there.

I think it just depends on the topology a little bit.

Even if it's a new PSU, if it uses an oldschool half-bridge design (which I know a lot of older CWT-built Antecs did), chances are you could probably get away with KMG or other GP 105C caps, simply because the switching frequency will likely be relatively low. Of course, there is a little bit more to it than that. (For example, the PSU transformer ratios can be designed to output higher voltage pulses with narrower width to get more "oomph" out of the switching transistors. This will put higher stress on the output caps and inductors. So that's one example where even a relatively low switching frequency PSU may benefit from low ESR caps. But even in those cases, GP caps with enough capacitance and adequate ripple current rating for the circuit should still handle the stress OK.)

[pc7fan]

They used KMG for a reason, those aren't areas that require high ripple current, thus 105° GP caps deem appropriate for decoupling. I have also seen good 10mm 16/3300 and 6.3/4700 Fuhjyyu TN that indeed tested around 3300 and 4700 respectively in old SP Antecs and Apevias. Since Fuhjyyu TN are general purpose, reducing to 3300u low-Z caps shouldn't be an issue whatsoever, plus that may reduce the risk of oscillations and will roughly match in ripple current ratings.

[mockingbird]

I think SMG is overall superior to KMG for general purpose applications. True, SMG is 85c and KMG is 105c, but IIRC, SMG is 2000h @ 85c while KMG is 1000h @ 105c. I use SMG in wall warts where I replace what is usually an old, crusty 80s Taiwanese cap.

[pc7fan]

Originally Posted by mockingbird

I think SMG is overall superior to KMG for general purpose applications. True, SMG is 85c and KMG is 105c, but IIRC, SMG is 2000h @ 85c while KMG is 1000h @ 105c. I use SMG in wall warts where I replace what is usually an old, crusty 80s Taiwanese cap.

SMG is a fine choice, but remember that estimated lifetime doubles each 10 degrees Celcius less. KMG then should technically have double the lifetime than SMG at 4000 hours at 85C, if other factors are the same. KMG is the high-temperature variant of SMG.

[mockingbird]

Originally Posted by pc7fan

SMG is a fine choice, but remember that estimated lifetime doubles each 10 degrees Celcius less. KMG then should technically have double the lifetime than SMG at 4000 hours at 85C, if other factors are the same. KMG is the high-temperature variant of SMG.

I disagree with this characterization. I think SMG uses better materials than KMG. I believe they do better in the long term (at standard temperatures that is).

The reason for my logic is that SMG has a higher ripple rating per given size. For example, for 12.5 x 20, KMG is rated at 920mA and SMG is rated at 1295mA.

I think as a whole KMG is better than Nichicon VR or VZ (decent series but with a finite lifetime), but still not as reliable as SMG. When I'm looking for caps to replace old 1970s BEG (butyl ethylene glycol) capacitors in equipment, I go to SMG. I think SMG could perform well for many decades, while I wouldn't trust KMG for more than 10 years.

[momaka]

Originally Posted by pc7fan

reducing to 3300u low-Z caps shouldn't be an issue whatsoever, plus that may reduce the risk of oscillations and will roughly match in ripple current ratings.

For PSUs, reducing capacitance while also reducing the ESR actually increases the risk of oscillation... but in most cases (it does vary a bit with designs) you have to get silly with the ESR reduction to cause that. Off top of my head, only Sirtec -made PSUs will start to oscillate badly if you use lower ESR caps than what they had (and in some cases, some of their designs oscillate at low and high loads by default.)
But by far and large, reducing capacitance (drastically - as in half or less) is much worse and more likely to cause oscillations than anything else.

On the other hand, increasing the ESR simply decreases ripple filtering. Again, you have to go silly with increasing the ESR to cause oscillations... and as long as you use the same capacity caps as the originals, there won't be a chance for the ESR to be so high as to cause oscillation. So because of that, you could technically use GP caps in any PSU. Of course, they may not last long and the ripple output may be higher... but again, it depends on the PSU design.

Originally Posted by pc7fan

SMG is a fine choice, but remember that estimated lifetime doubles each 10 degrees Celcius less. KMG then should technically have double the lifetime than SMG at 4000 hours at 85C, if other factors are the same. KMG is the high-temperature variant of SMG.

Exactly.
So it comes down to application, really.

SME does indeed have higher rated ripple current, though, as mockingbird pointed out. Thus, at 120 Hz rectified DC applications (such as the output of line-connected transformers), it should technically be able to take more abuse than KMG.

On the other hand, in applications where you don't really expect a lot of ripple current but just need a GP filter (such as 3.3V, 5V, and 12V rail on the PCI slots, USB connectors, and Firewire I/O on motherboards), KMG may be a better choice than SMG, simply due to the higher temperature rating. This is why on the Xbox 360, Hann-Star used KMG (or Nichicon VZ) for the 16V, 100 uF caps instead of 85C series. Those Xbox 360 consoles can get pretty toasty inside. That said, I have not seen any of those KMG or VZ small caps fail inside them. So I think KMG is an OK series.

BTW, KMG and SMG are the successors (and lead-free versions, if I'm not mistaken) of KME and SME series, respectively.

[Topcat]

Originally Posted by mockingbird

I'm still in touch with Joe on occasion.

Not sure how I missed this one...but I'm glad to hear that! The way he just 'vanished' had me worried that his government just 'disappeared' him, I'm told is a common thing over there. Give him my regards next time you're in touch with him.

[Paar]

I'm currently in a need of 4700uF 10mm caps and noticed that Rubycon ZLQ series is suitable for this. Mouser has them and Digikey too but the postage price is crazy, at least for an individual person like me who just wants to repair single PSU. Are there any other sources I could use?

[pfrcom]

https://badcaps.net/forum/member.php?u=15285

He's in Prague