Hi:
Why do the datasheets always include capacitors in the voltage pins (see enclosed image)???
Those are "decoupling" capacitors trying to keep the switching noise of the chip out of the power rails and keep noise on the power rails out of the chip.
johnwasser:
Those are "decoupling" capacitors trying to keep the switching noise of the chip out of the power rails and keep noise on the power rails out of the chip.
and what would be the consequences of not using them? are they really necessary?
If you leave them out the circuit will probably work most of the time. If you want the circuit to work more reliably, leave them in.
Gilgamesh90:
and what would be the consequences of not using them? are they really necessary?
Random pattern-sensitive failure. Basically anything and it would drive you mad
trying to figure it out.
Always include adequate decoupling in any circuit using high speed logic chips,
without exception, and your life will have fewer nasty surprises.
Fast logic circuits generate voltage swings of billions of volts a second and
current changes of millions of amps a second - at these rates stray inductance
of the power wiring is significant and you have to provide a low inductance
path from the power pins to enough capacitance to source / sink the current
changes. This means ceramic capacitors very close to the chip. If you have
a very fast 'scope and a low impedance probe you can see the spikes/dips in supply
voltage on the nano-second timescale if decoupling is inadequate.
Thanks for the info! If one is using the L293, is it suitable to use decoupling capacitors, even they are not shown on the data sheet???, the author of this article:
http://letsmakerobots.com/content/control-your-motors-l293d?page=1
he advised that the schematic was based on the ladyada motor shield.
It's always suitable to use decoupling capacitors where there's a requirement for clean power. That's especially true with a device like a motor driver where heavy on/off current draws are occurring.
Looking at it the other way, your chances of creating a problem by adding a decoupling capacitor is extremely tiny compared to the number of problems that you'll avoid (emphasis on MarkT's "drive you mad" comment). Start with the capacitor, test your circuit, then remove it later if you're absolutely certain it's not needed.