Are the exact values of power supply filter caps important?

[Nitefly182]

Im building a pedal with some power filtering and for some reason Im out of 100uf caps except in sizes way too big to be reasonable. I assume I can use something like a 220uf and not have any issues right? Could I use a smaller cap like a 47uf and still be OK?

[km-r]

its ok to use larger supply decoupling caps...
its also recommended btw... to reduce hum and nasty things...

in tube amps, the supply caps vary the response of a crankin tube amp...

[dano12]

Someone posted (DougH maybe) a while ago that one of the trade secrets of power supply filtering was to use several smaller value caps instead of one.

I haven't been able to track that down in terms of theory, but your 220uf should work fine.

[R.G.]

You could try reading "Power Supplies Basics" at Geofex: http://www.geofex.com.

The short answer is that at power line frequencies, filtering caps can be larger and be fine; and you can use multiples in parallel just fine.

For switching power supplies, the high frequency effects of the capacitor output come into play. In that case, you need to worry about the Equivalent Series Resistance, Equivalent Series Inductance per capacitor, and the inductance of the wiring.

In those cases, you will need either special low-ESR/ESL caps or an array of smaller value caps in parallel to run the total ESR/ESL down, or maybe even conductive plane hookups to lower the interconnection inductance between caps.

At very high frequencies, you'll need both arrays of parallel caps and decreasing-value caps to lower the parasitics of the caps as well as low-value, low-ESR ceramic caps right at the load.

[Sir H C]

You can usually safely double the cap value, but don't go crazy or you could overstress other parts such as the rectifying diodes. 

[R.G.]

That was a critical issue in the days of vacuum tube rectifiers, but it's much less so today.

What Chris is referring to is the peak currents caused by very big capacitors. If you read "Power Supplies Basics" at GEO, you will see that the current into the caps all flows in one pulse right at the peak of the AC power line sine wave. This peak is many times the DC average current out of the power supply.

Solid state diodes are capable of huge peaks in current and in general; as a practical matter it's hard to afford enough caps to make the caps too big for diodes that are intended as power supply diodes, from the 1N400x series on up.

But he's right that it is possible to get enough capacitance - or more properly, enough lowering of ESR/ESL - to get current peaks that will damage diodes.

[frank_p]

I have red that 2200uF by Amperes before the regulator and 10% of the first value after the regulator.
Is this a good rule of thumb ?

In French and at the bottom:
Ref:
http://www.projetg5.com/modules.php?name=Tutoriaux&rop=tutoriaux&did=36