Capacitor identification

[trodas]

Guys, I have my ideas about what it is, but to be sure, I need verification. So, the capacitor is a small SMD suxxka, d6.3 for the round part and it looks like this one:



However on the cap is with black print (the negative is also black as on pic) this:

t3
100C
UD

The card is relatively older (2005 or 2006) and I measured the voltage on the cap already (3.4V), so I need more or less confirmation on the capacitance. My quess is obvious (100uF), but someone can back my quesses with facts?

[Pentium4]

Re: Capacitor identification

maybe nichicon UD series?

[mariushm]

Re: Capacitor identification

Nichicon UD 100uF 16v


6.3 10 16 25 35 50
j....A....C..E...V... H

[trodas]

Re: Capacitor identification

Pentium4 - mariushm - thank you very much, guys! You seems to be correct. Nichicon UD datasheet confirm this:


And Digikey.com agree:
http://www.digikey.com/product-detai...62-1-ND/590237

Good to know that I was not mistaken. The "t3" worry me a bit, but documentation say, that this is just a Lot No., so probably nothing to worry about it.


Any ideas what similary looking SMD cap, but just d4 is? On it are:
10
16s.
005

10uF 16V...? Or 16uF 10V?

That one:

[trodas]

Re: Capacitor identification

Bump, pls?

[mariushm]

Re: Capacitor identification

Capacitors are pretty much made with standard values. 16uF isn't a standard value for capacitance, but it is one for voltage.
So it makes most sense to be 10uF 16v.

Though, that capacitor is kind of big for a 10uF 16v capacitor.

Why are you messing with it anyway?

[PeteS in CA]

Re: Capacitor identification

Concur with mariushm. 10uF, 15uF, and 22uF are fairly common standard values, but not 16uF. Compared to the nearby SOT-23-style and D-PAK devices, I'm not sure that cap is large for 10uF, 16V, though.

[trodas]

Re: Capacitor identification

Well, it is a d4 in diamater, so in reality it is pretty small. Yet still I thinking about bumping the capacitance to 22uF - a nice SMD ceramic cap can easily are todays 22uF - even smaller 1210 size that I wanted it to be (1812, 1808)...

And I believe that this cap might be responsible for some of my stability problems, as there is a 1.55V (AGP voltage) on it...

[mariushm]

Re: Capacitor identification

I wouldn't recommend using a ceramic capacitor there. High capacitance SMD capacitors have... problems.
If you really want to, I'd suggest a polymer capacitor, I guess these would work:

http://www.digikey.com/product-detai...3-1-ND/3465979
http://www.digikey.com/product-detai...9-1-ND/2327075

[trodas]

Re: Capacitor identification

Well, I have the very same feeling, that it might not be best idea, because the ceramic cap characteristic are way off the classic elyte cap.

However there is a problem. I would use polymer, BUT it must be d4 ... your tips are d5 and with all the help and skill - they cannot fit there. Guaranted. I will have to even desolder the close cap to get desoldred the old and soldered new cap.
There is simply no way to fit a d5 there. Believe me. I would love to, but I just... can't.

What problems plague these hi-capacitance ceramics anyway? There are even 47uF 6.3V ceramic caps:
http://www.digikey.com/product-detai...6-1-ND/4157642

I was under the impression that these would be beneficial for example right under the CPU, where is place for lot SMD ceramic caps and the place is... empty:

[PeteS in CA]

Re: Capacitor identification

Potential problems with SMT high value ceramics:

1. Temperature coefficient;

2. Voltage coefficient;

3. Cracking with larger parts.

So, the part you linked: X7R is OK, the "worst" tempco that should be considered for GP usage; 1210 is probably OK for susceptibility to cracking; 6.3V ... to avoid the voltage coefficient becoming relevant, the part really should be rated for at least 10X the circuit DC voltage. Now, a 22uF, 16V (or 25V), 1210, X7R might be OK (I have a hard time saying that, I have such a visceral dislike of SMT ceramic caps).

[trodas]

Re: Capacitor identification

The AGP voltage can be adjusted in BIOS between 1.50 to 1.90V. Default being 1.55V and I find it to be the most stable. It is only 4x lower that the 6.3V the cap is rated for, but what options with the d4 size do I have?

[trodas]

Re: Capacitor identification

Actually, looks like there is a way. Nichicon ZR series are very small:

...and they have (the smallest) d4 SMD caps starting from 33uF 4V, to 47uF 4V and 22uF na 6.3V ... and even 10uF 16V in the d4 SMD...!

The ripple is laughably low 28mA:
http://www.digikey.com/product-detai...5-1-ND/3983940
...but maybe this is better that the 47uF 6.3V ceremic SMD cap?
I don't know, but I do know that X75 ceramic cap can go to +125°C and the Nichicon ZR is only +85°C ...

What make me wonder is also how much ripple curent the ceramic cap can deliver. I would be inclined to think that "rather a lot", but there is on the very detailed Taiyo Yuden datasheet only information about "Charge/discharge current: 50mA max." - and that it is. This is still considerably more that the 28mA, but it is that the maximum ripple current? And if it is, is not this a bit low ripple anyway?

[mariushm]

Re: Capacitor identification

Bend the leads a bit and you should be able to pull out the brown rubber from the smd capacitor bottom, then solder the leads directly on the pads as if it was a radial capacitor.

A D5 would then fit just fine in that place.

[PeteS in CA]

Re: Capacitor identification

Is the circuit application power filtering - i.e.requiring low Z and high ripple current capability? Or GP local decoupling?

[trodas]

Re: Capacitor identification

mariushm - I could do that, but is there enough space up for that? I worry about it a lot. Better to have some options then...


PeteS in CA - no idea. How should I know...? All I know that just the same 1.55V is on the near smaller chip with there legs and cooling back soldered to the mobo. This is on it:
LIO37G
QHE11
Is this some power switching part?

[trodas]

Re: Capacitor identification

BUMP for some suggestions. As the crashes are becoming more and more frequent I consider a ghettom mod - just soldering there a normal Panny FM 120uF 6.3V cap...

For longer time - what about some tantal caps if the ceramics should be too much for the circuit? What is a good, recommended company that producing tantalum caps (witch are available at Digikey...) ...?