I usually see that smoothing or filter capacitors used after rectified mains, consist of two, or more capacitors in parallel. For example, it's common in treadmills to find two 1500 uF capacitors in parallel instead of using just one of 3000uF or 3300uF (more common). Is there any advantage of this when it comes to electrical/electronics issues? Is that a cost issue?

 

You often can achieve higher ripple current rating and lower ESR by using multiple capacitors in parallel rather than a single cap of the same total capacitance and voltage rating. Improving these ratings translates to longer lifetime. The cost is likely to be a bit higher using multiple caps, but not always. Sometimes you simply can't get the rating you require in a single capacitor. Sometimes it is done for physical shape reasons, for example using multiple short capacitors when the only equivalent as a single is too tall to fit, though a short "fat" capacitor typically has somewhat higher ESR and lower ripple current rating than a cap in the same series, capacitance and voltage that is tall and slender. Sometimes a single big "snap-in" capacitor is preferable to multiples that have ordinary wire leads because it more secure on the circuit board.

In the past 15 years or so it has become very much easier to buy good quality electrolytic capacitors off the shelf, so much so that it can be quite time consuming to pick the best combination of characteristics and price to suit your require.

 

There can be a few reasons for using multiple caps in parallel.

If the caps are the same size, then it might be to lower the effective series resistance and/or inductance of the effective capacitance. It might also be to distribute the total capacitance around the circuit so that the the charge storage is closer to where it needs to be used. It might also be to allow physically smaller capacitors to be used in order to meet form factor requirements.

If the caps are of significantly different sizes (on order of magnitude or more) then the likely reason is that the capacitors have what is known as a self-resonant frequency and, above that frequency, the parasitic inductance dominates and the device looks like an inductor and not a capacitor. The larger the capacitor, the larger the parasitic inductance (in general) and so the lower the self-resonant frequency. So in order to get proper bypassing of the supply at high frequencies, you need to use small capacitors. But this is generally okay because the magnitudes and durations of the signal transients get smaller at higher frequencies and so a smaller capacitor can suffice. But at the lower frequencies the charge storage needed to deal with the noise are higher and so higher-valued capacitors are needed.

A common strategy for good bypassing (often times too good, don't do more than is needed, including a safety margin) is to put some large capacitors at the power entry point. Then put some moderately sized bulk-storage capacitors scattered around the circuit. Finally, put some small-valued capacitors as close to the supply pins of every IC (or subcircuit if a discrete design). This is done in conjunction with good power supply routing and signal routing, possibly including the use of stitching capacitors and inductors (such as ferrite beads) to minimize and channel noise signals away from sensitive portions of the circuit.

 

And in bulk caps, like some electrolytics, where their chemical dielectric layer
forming "juice" can dry out still leaves, potentially, other non leaking cap
functioning, so limited redundancy.

That of course is if cap does not short due to increasing V in circuit where Ceff.
is dropping.

Regards, Dana.

 

Cost, size, footprint, availability.

 

Forgot to mention: Sometimes two capacitors are used to support 120/240 volt operation. Typically there will be either a movable jumper wire or a switch to set this. At 240 VAC the caps are in series on the output of a full-wave bridge rectifier. At 120 VAC the configuration is as a voltage doubler.

explained here (don't know how well - I just searched for a schematic): http://www.industrial-electronics.com/switching-power-supply_1-6.html