[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: Edge ringing filtering



From: Dan McMahill <dan@xxxxxxxxxxxx>
Subject: Re: gEDA-user: Edge ringing filtering
Date: Thu, 02 Mar 2006 18:21:40 -0500
Message-ID: <44077E04.402@xxxxxxxxxxxx>

Hi Dan,

> > Also, 100 nF caps may be beyond self resonance at 200 MHz depending on 
> > their construction.  Check the manufacturer's data sheet.  It may help 
> > to put a smaller cap (100 pF) in parallel with the 10 & 100 nF if 200 
> > MHz is your problem.
> 
> You know... I don't completely buy this argument anymore.  What do I 
> care if my bypass cap is beyond self resonance?  What I care about is if 
> the impedance to ground that it presents is very low.

This is indeed true. A bypass capacitor is a shunt which is intended to have a
lower impedance over some range of frequency than the load. The effect is that
most of the current in those frequencies goes down the shunt, reducing the
variation of voltage (in time) since the offending energy (voltage in
frequency) is mostly shortend down. The bypass capacitor is in that context a
parallel/shunting voltage regulator.

>  So pretend I have 
> 2 caps that are exactly modeled by a series RLC network and I keep R and 
> L constant.  4x the C gives 1/2 the self resonant freq, but except for 
> right near the resonance of the 1X cap, the 4x cap has a lower 
> impedance.  At low frequencies, it has 1/4 the impedance and at higher 
> frequencies it is the same.  So my take on the "put smaller caps in 
> parallel" old wisdom is that you have to make sure you're really getting 
> a benefit.  The last I looked at some low voltage 0603 (or were they 
> 0402) ceramics, the data seemed to suggest that bigger was always better.
> 
> I think the place where you can get a benefit is when your bigger caps 
> put you into a different package or technology that caused significant 
> difference in R and L or the required board parasitics to handle the big 
> package.

Some writers have proposed that you need a difference of about two decades
between the caps to get the desired effect. Otherwise they even up too much
where as when you separate them enought in self-resonance frequency you get
two dips in impedance which work together.

Cheers,
Magnus