To decouple or not

Thread Starter

morganism

Joined Apr 13, 2012
4
After lusting after the Gecko EFM prototyping board, saw some articles in the DigiKey catalog talking about "sprinkling" caps all over your designs.

It was pointing out how big a power draw it was to do that, basically cutting battery life in half.

Where is the tradeoff point for de-bouncing, and quick to sleep and rise ?
 

chrisw1990

Joined Oct 22, 2011
551
i have always taken the general practice to decouple often..
its not so much noise, but about the power draw from items like uC's etc..
the general rule i apply is such:
near the psu, apply a larger cap, lets say 10uF, and next to an IC.. of small proportion, place 10n or 100nF.. for uCs id lean more towards 100nup to 1u. though they usually require multiple caps anyway so they all add up.
power draw isnt really an issue..
the minute value to which inrush current is increased.. meh, and i personally wouldnt run anything directly from a battery.. maybe a torch, but i buy them, dont design them..;)
 

Ron H

Joined Apr 14, 2005
7,063
After lusting after the Gecko EFM prototyping board, saw some articles in the DigiKey catalog talking about "sprinkling" caps all over your designs.

It was pointing out how big a power draw it was to do that, basically cutting battery life in half.

Where is the tradeoff point for de-bouncing, and quick to sleep and rise ?
Please provide a link to that article. I can't believe Digikey would make that statement.
 

Adjuster

Joined Dec 26, 2010
2,148
Large de-bouncing capacitors on lines that were switched frequently enough could conceivably add significant power consumption to a system.

Generally however, supply decoupling capacitors only require an inrush of current at power-up,and unless powering up and down is extremely frequent this is normally acceptable. These capacitors are generally not optional, as it is otherwise difficult to obtain the low power supply impedances required to keep rail voltages steady in face of rapid current variations.
 
Top