> > If you're going to be in electronics for a while, buy yourself a reel > > of 10k resistors and a reel of 0.1uF ceramic caps. > Nowadays I'd get 1uF instead of 100nF. > some people have forgotten to re-evaluate the assumptions when > the available parts changed the tradeoffs. > However, unless you routinely to RF, these kinds of things > come up rarely in microcontroller circuits. Others argue equally convincingly, perhaps, for reducing capacitance due to changing factors. "Modern" processor, and other, circuits have switching speeds many times greater than was the case when 0.1 uF's were first mooted the standard for general purpose decoupling. As correctly noted, the characteristics of 1 uF caps may exceed the capabilities of 0.1 uF's from "long ago". But the same applies to modern 0.1 uF caps and say 0.01 uF caps may have a vastly better response at 10's to 100's of MHz. While clock and port signal edges are purposefully slew rate limited ij many designs, a "square" edge still produces frequency components to many time the fundamental frequency. A 50 Mhz digital signal can, under the right conditions, produce "useful" spuries to 100's of MHz. A device drawing 10 mA at 3v3 has a notional DC load impedance of V/I =3D 3= 30 ohms At 10 Mhz a 0.01 uF cap has an impedance of under 2 ohms. At 100 Mhz under 0.2 ohm. A 1 uF will notionally be 1/100th of that so notionally more useful But both are far below the effective impedance of the device proper, and the 100 x decrease in impedance is liable to be illusory when lead inductance and other secondary factors are considered. For general low power switching noise in high speed circuits the 0.01 uF is liable to be as good or better. Where the 1 uF will come into its own is with "power switching" where significant energy is purposefully being switched, either at high speed or in transition events. Examples might be RF output stages (DC load variation rather than the RF component), USB load switching, smps transients as significant loads are switched etc. As Olin suggested, selection to suit the purpose is still wise, even in the world of capacitors better than your father even imagined. But, in demanding applications your grandfarher's solution may still be best - 2 or 3 caps of widely varying value placed together at key locations. In some cases a 1 uF and 0.01 uF may work magic. In others even a 1 uF and 1 NF or even 10 uF and 1 NF. Good old days practice was often 3 caps with one in the 100's of pF to low NF range, something one or two decades greater and an ecap or large non-polar around the uF range. still a goof idea [tm] in some cases - even non-RF. eg a 12 bit or better ADC will often have great trouble achieving its potential accuracy without this sort of care and more. A 20+ bit Sigma Delta almost requires arcane knowledge and an act of parliament to work properly. Russell --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .