Could you explain the reasoning for the cap change?
The 0.1uF capacitor is the typical size recommended in this
application. Some applications could require more some less. The
downside of larger ceramic capacitors would primarily be size and
cost.
If you want to quantify the value you could (1) measure the voltage
ripple in your application or (2) estimate the ripple using the
formula I = C dv/dt. You would need to determine the switching currents.
0.1uF is a good comprimise of capacitance/size/cost. Since this local
filtering the placement is almost as critical as the value. Keep this
capacitor close to the uC and keep the pin length short.
If the MCU is destined for it's own board, I would most certainly add
onboard regulation. However, if it is just part of a larger board
(say a led matrix with half a dozen ic's, resistors, etc, etc.) Would
I be better off with multiple power supplies (possibly driven from a
single input) or would a single regulated supply (of sufficient
current) be suitable?
Unless the microcontroller is switching large loads (which usually
isn't the case) I would usually not add a separate regulator. I have
built systems where the unregulated bus was distributed and local
regulation was used for different sections.
If the distance between the uC and regulator is large you may want to
add some larger filter capacitors near the uC. I place the 0.1uF
withing a 100-200mils of the uC. The larger filter caps within a few
inches or so.
On my boards like the ZB1 and NBx I have output pins for the
regulated voltage and the unregulated bus. This enables each
application to provide additional voltage regulation.
The schematics are in the datasheet.
(mental side track: assuming a single power supply, would the
use of optocouplers (or similar) be required to keep things
"separate"?
If you are interfacing to line voltages you should have isolation.
You should be powering from an isolated supply (like a wall-wart)
and be using solid-state or mechanical relays. Almost all solid-state
relays have internal optical isolation. Mechanical relays are isolated
by the driving coil.
(* jcl *)
www: http://www.wiblocks.com
twitter: x.com
blog: http://luciani.org