Your basic theory is sound... to a point... but unfortunately, it's a bit detached from the reality of current PSU design.
Ripple reduction is achieved with a circuit consisting of a series impedence, which can be a linear resistor (R), which has a resistance (in ohms - Ω
that remains relatively constant with frequency, or an inductor (in henrys - L), which has a resistance that rises with frequency, and a shunt capacitor, which has a resistance that decreaes with frequency. Any D.C. voltage dropeed across the series impedence is wasted as heat. Indutors are preferred because all of the resistance, and thus, the loss, is much lower at D.C.
The frequency is calcuated to the point where power is reduced by half (-3 dB) by the equations:
f = 1/2 x π x R x C for a resistor and
f = 1/2 x π x L x C for an inductor
Assuming a PSU of competent design, a typical packaged PSU already has a lot of capacitance on each ouput. Depending on the design of the PSU and the specifications of the components, the effect of adding more capacitance could ramge from almost nothing to catastrophic failure. Consider three conditions:
1. In combination with the series resistance, the capacitance is too low to add significant further ripple rejection, in which case you'll get little if any measurable improvement.
2. The capacitance is sufficient to improve ripple rejection, which would probably require very large capactors to come close to the amount of capacitance already in the unit, and you'll run into space and cost limitations.
3. The effective impedence of a capacitor falls to near zero ohms and acts as a short circuit to ground at high frequencies. Adding enough capacitance across an output to have a significant effect on ripple could cause the current limit protection circuitry to see the capacitor as a short circuit and shut down the entire supply. If the current limit protection circuitry is inadequate or if it fails, it could cause a catastrophic failure. You'll know if you smell smoke or see flames.
There are many other factors in PSU design including the non-ideal characteristics of real world components and the effects of active ripple rejection circuitry, which can accomplish far more ripple reduction than is possible with a passive L-C or R-C network, but without knowing more about the actual circuitry, simply hanging a lot of capacitance on the output of an existing design is at least speculative and at worst, catastrophic.
Facebook
Twitter