another quesiton on power supply caps...

[bosone]

this one is very basical...

from the lessons i took at university (a lot of years ago!), i remember that putting two caps in parallel is equal to having only one cap with a capacity that is the sum of the two.

so: why do you put two caps before and after the 78L05/09/12 regulator?

a 10uF one (el.cap) and a smaller 100nF = 10.1 uF. is the 0.1 uF absolutely necessary?

since it can be used also an el.cap. with a value higher that the one stated (10uF), if i mount a 22uF or more el.cap, why not to remove the smaller 100nF one?

[smashtv]

The quick answer:

Both are for filtering noise out of the power, and each value catches a different type of noise, better than a single capacitor can.

In this case, the 10uF cap serves to catch some AC ripple and noise (and harmonics of that noise) caused by the regulator.

The .1uF serves to catch switching noise caused by the IC's (to help keep this noise from showing up on everything else on the power rail).

Even though the output of the regulator is linear (way less noise than a switching power supply design like in your PC), it is most certainly not noise free..... :)

Long answer:

In -most- designs you will see the .1uF cap next to each chip, but since the CORE only has one IC that generates switch noise C3 can effectively catch it before it shows up in the power to your other modules.

Before it is asked: Why don't the modules need a .1uF cap next to each chip?  HC class chips have a higher power noise immunity than their standard counterparts, and with unused inputs pulled up or down, there is nothing sensitive enough to be affected by the switching noise.

Best!

Smash

[moebius]

i remember that putting two caps in parallel is equal to having only one cap with a capacity that is the sum of the two.

That's correct.

a 10uF one (el.cap) and a smaller 100nF = 10.1 uF. is the 0.1 uF absolutely necessary?

Well, do you remember equation for time constant in Cap circuits?

The bigger cap is used as storage, that "slowly" charges and discharges as voltage changes, it keeps voltage pretty stable. Its "size" isn't that critical, as long as it's there. This is especially important after rectifier, which outputs pulsating DC.

The smaller cap has significantly smaller time constant <-> it charges and discharges in much faster rate <- > it reacts faster on voltage changes <-> It takes care of the "ripple voltage" left in powerline after the bigger cap. ;)

Omitting power supply bypass caps is not a smart or healthy thing to do ;)

Bye, Moebius

[bosone]

i'm working on my 2xsid+1xcore.

i decided to build a "power supply module" which has the rectifier, several regulators (05/09/12V), and the two caps after the rectifier.

since on every PCB (sid/core) there are the big elcap (10uF) and the small cap (100 nF), i was wonder if, on the power supply modure, i have to mount also another 2 caps (10uF+100nF) on each power output, before the connection that will bring the power to each module.

since i have at home a lot bigger el.cap. (so it came the first question about the elcap.on power supply) but no small caps, i was wonder if it could be safe to remove the small ones from the power supply (here it came the second question, maybe i should have adressed the subject more clearly from the beginning...), since they are also on the pcb...

[Duggle]

quick answer is no, the monolithic ceramics mounted *close* across the output of voltage regulator ICs perform filtering at very high frequencies which is essential for long term stability of the voltage regulator.

Electrolytic capacitors "dry out" over time (measured in months and years) resulting deteriorated high frequency performance (although they can still hold a charge and are still useful for low frequencies).

Yes, parallel "ideal" capacitors add, but real capacitors are never "ideal" and the importance of resistance, inductance and leakage resistance depends on the specific application.