component layouts?

[Harry Bissell]

Hi Ian (and all)

Good point about ferrite beads. They are very misunderstood, but work well
in a lot of situations. Ferrite beads are about useless at audio frequencies...
but
perform very well at RF...   High frequencies do not use the center of the
copper
at all, they ride on the surface of the traces (and especially on the edges)...
The
ferrite beads act as little sponges (they add lossy impedance) that suck off
those
highs and literally "hang them out to dry".   Now if the problems are from
voltage
spikes, the fast rising edges of the spikes will contain many high harmonics,
which can
slip past some decoupling caps.  The idea is to have the filter work at all
frequencies.
So a big cap (like 1-10uF Tantalum or Aluminum Electrolytic) with good low
frequency response, and a small cap with good high frequency response (like a
.01-.1uF ceramic)
is common.  Keeping inductances down helps, as does the ferrite beads. The beads
are
most effective when put very close to where they are needed.  My latest fuzz-box
design
has two, one on the resistor lead located right at the input pin of a CA3080,
and the other
right at the Gm bias pin. This lets it work well (no losses) at audio
frequencies, but kills off
the radio reception.

Care must be taken with using combinations of capacitors and inductors... it is
possible
to make combinations that actually resonate and work worse than just caps alone.
Resistors as series elements can work just as well, but the voltage drops might
hurt here...

There is a guy who writes for Electronic Design, think its Ron Mancini... he
writes about
signal integrity issues. Usually advanced topics, but sometimes interesting.
Maybe search
those sites looking for words like "decoupling", "power distribution", etc...

H^) harry

Ian Fritz wrote:

> Good suggestions from Jules and Jim. I would add a suggestion from recent
> experiences I have had working on VCO design -- include lots of ferrite
> beads on the power supply lines. I divided the circuit into several circuit
> blocks and ran separate power lines for each section through ferrite beads
> back to a common supply input. This is the first time I have been able to
> virtually eliminate switching spikes from showing up in the expo converter,
> etc. This may be overkill for many applications, of course, but for critical
> circuits with wide ranging voltages it really works well. It has a big
> advantage over the usual technique of adding series decoupling resistors in
> that the supply remains "hard" at dc, in other words, there is no shift in
> supply voltage as the load current varies.
>
>   Ian
>
> ----- Original Message -----
> From: "Jules Ryckebusch" <ryckebu at ersg.san.mrms.navy.mil>
> To: "Jim Patchell" <patchell at silcom.com>
> Cc: "zapsn0n" <n0nspaz at loa.com>; <synth-diy at node12b53.a2000.nl>
> Sent: Tuesday, September 26, 2000 2:02 PM
> Subject: Re: component layouts?
>
> > Ohhh Doggie! Compnent layout grounding and bypassing is always an art
> kinda
> > thing. There are some good downloadable things on it from Analog Devices
> web
> > site in their ap notes. I have found that grounding is usually the cuplrit
> in
> > most cross talk issues but being neat counts the most. I usually build
> "one
> > of's" that are bread boarded. I keep the power supply leads as short as
> > possible and bypass all IC's at the pins with a 10-47uf electrolytic and a
> .1uf
> > poly or ceramic cap too. On a couple occasions (Power amplifiers) I have
> found
> > the need to add additional bypass caps with in the circuit in places where
> a
> > current sink/source that was stable was needed. Another helpful thing is
> to
> > mount the PC board close to a ground plane. ie metal chassis of case etc.
> >
> > Jules Ryckebusch
> >
> > Jim Patchell wrote:
> >
> > >     Boy, talk about opening a can of worms.....
> > >
> > >     Bypassing is both art and science.  My rule of thumb 1st of all, 1
> > > bypass cap, to ground, for each power supply pin.  There are exceptions
> to
> > > that rule of thumb, but that is another story.  It is also best to keep
> the
> > > distance between the bypass cap and the power supply pins as short as
> > > posible.  The idea of the bypass cap is to provide current for those
> peak
> > > useage times.  For analog sutff, this is not quite as critical as for
> high
> > > frequency switching, but the idea is still the same.
> > >
> > >     Another general rule, I cannot think of a case where it is posible
> to
> > > have too much bypassing.
> > >
> > >     Now, cross talk is another problem.  Even with good bypassing, you
> are
> > > going to want to look at the circuit and ask, where can extraneous
> signals
> > > get into things.
> > >
> > >     In conclusion, all I can say is that this is a pretty big subject.
> I
> > > don't know of any articles on the net that cover this.  You will
> probably
> > > get enough replys to make an article though1 :-)
> > >
> > >     -Jim
> > >
> > > zapsn0n wrote:
> > >
> > > > hi all,
> > > >
> > > > can anyone point me to some general online info & tips on component
> > > > layouts, e.g. the placement of capacitors on DIY circuit boards to
> avoid
> > > > crosstalk, etc etc  ?
> > > >
> > > > thanks,
> > > > mark scetta
> >
> >