Decoupling capacitors revisited

[Dicky Mint]

Hi Just a quick question on decoupling voltage regulators.

The datasheet for SMD 7805 shows a 330nF resistor across the input to ground.

As SMD 1uF are easier and cheaper to find than 330nF capacitors I'd like to use them to decouple my SMD voltage regulator.

Is this folly? Or is it just interpretation?

I want to use the voltage regulator to power a circuit which should take just a few milliamps.

Rick

 

[srnet]

For a regulator, the capacitor quoted is often as a minimum, what does the data sheet say ?.

In 1206 SMD, you can get 1uF or 0.33nF easily enough. The 1uF might be cheaper, and its what I would use.

 

[Dicky Mint]

The datasheet doesn't seem to elaborate a great deal it only gives the values as
330nF (input) and 100nF (output) in circuit diagrams!

I was a bit worried that perhaps there was an important reason why 330nF and not 1uF?

I couldn't think of one but that's why I'm asking?

 

[JimPerry]

In general capacitors are +50 - 20% and (except for critical timing circuits) generally "close enough" - unlike voltages (9V WILL fry your Picaxe). I've yet to hear of an incorrect capacitor value frying a circuit - OK electrolytics back-to front tend to explode ...

 

[russbow]

They are NanoFarads, not MicroFarads, a factor of 1000 less. I suspect it's to do with preventing parasitic oscilations or whatever.
Certainly not to be confused with smoothing / reservoir capabilities.

 

[g6ejd]

Most of the choices (now largely historical) were due to inductance of the capacitor construction, for example many layers of foil were inductive and hence choosing a small value of sub 1uF usually ensured the desired construction - of minimum inductance.

A more exact value of a decoupling capacitor can be calculated if the load variation is stable and known, then from energy required in a given time, the value can be deduced.