Shuttle FX41 Motherboard

[dracobuild]

This board has the following capacitors:
6 x 3300uf 10V 10mm
6 x 1500uf 10V 10mm
9 x 1000uf 10V 8mm
4 x 1500uf 16V 10mm

I have been able to locate quality replacement caps for everyone except the 3300uf 10V in 10mm size.

[willawake]

you wont find less than 12.5mm on the 3300 10v

[kc8adu]

use 6.3v caps
they used 10v probably because they were cheap.
the 16v ones are on the +12 input to the vrm so they must be 16v caps
all the others can be 6.3

[Neo2_000]

On many boards I often see smaller value caps rated at 16 or even 25 volts in other areas of the board beyond the CPU. Is this because those values just happen to be manufactured in those particular voltages for reasons of versatility or is there a real need for these higher voltage caps on the motherboard?

[kc8adu]

often its due to savings by buying a large quantity of one value vs several different voltages.
not to mention simplifying insertion machine setup.

[willawake]

use 6.3v caps
they used 10v probably because they were cheap.
the 16v ones are on the +12 input to the vrm so they must be 16v caps
all the others can be 6.3

is it possible to use 10V but with higher uf than previously used on the +12 input for the vrm....or not recommended?

[Neo2_000]

So the motherboard voltages in these low value cap positions are considerably less than the rated cap voltage under all conceivable operating conditions?

[AK0R]

is it possible to use 10V but with higher uf than previously used on the +12 input for the vrm....or not recommended?

I certainly wouldn't recommend it. The capacitance is important for voltage smoothing, but the voltage rating is important to prevent perforation of the capacitor film (paper, mylar, foil, etc.) due to internal arcing. Capacitors should *NEVER* be installed that have a lower voltage rating than what is actually present in the circuit. That doesn't mean you can't, or shouldn't, install a lower-rated capacitor than what was there originally. There have been several threads posted here that indicate 10V capacitors were replaced with 6.3V units, but that was on Vcore, which never gets above a couple of volts. A good rule of thumb in the radio world has been twice the actual power dissipation for resistors, and twice the actual voltage for capacitors. Motherboard manufacturers have reduced these margins of safety in the interests of cost savings and board space to almost nothing. Capacitor manufacturer curves show substantial de-rating as actual voltage exceeds capacitor rated voltage.

[willawake]

good info, thanks AK0R

[AK0R]

So the motherboard voltages in these low value cap positions are considerably less than the rated cap voltage under all conceivable operating conditions?

Not necessarily "considerably less", but for low-value capacitors, it comes down to making a package that is large enough for a human to handle, and the consequential voltage rating for such low capacitance. An easy way to lower the capacitance (in the manufacturing process) is to increase the spacing between the layers of film, which also raises its voltage rating. Typical of most capacitor manufacturers, Rubycon doesn't make a 1μF radial lead aluminum electroytic capacitor lower than 50V. It also may be that a "proper" voltage capacitor doesn't have the desired properties. And as has been mentioned elsewhere, economy plays a tremendous part. If one rating fits several different requirements, a manufacturer isn't going to stock several different voltage ratings of a particular capacitance value.

[MD Willington]

often its due to savings by buying a large quantity of one value vs several different voltages.
not to mention simplifying insertion machine setup.

Bingo...they don't have to load up reels of a different value and it simplifies the masking used in the chip shooter...

but it's not like a chipshooter is a slow device anyway...

[Neo2_000]

Interesting stuff. Thanks Akor.

What is your opinion then on the wisdom of replacing for example a 470uF x 25V or 470uF x 16V with a lesser rated cap on the motherboard.

Maybe I should ark up a board and do some measurements but I would be concerned that I might make a mistake by taking measurements when the particular circuit involved is not under full load.

Neo.

[kc8adu]

i wouldnt try it unless you know for sure what the cap will see in circut.
like the vrm caps at 10v when there should never be more than 2v.
thats why i say use 6.3v rated caps there.
if that section sees more than 6.3v the caps blowing their tops its the least of your worries ;-)