Re: [SI-LIST] : bypass cap question (long, simple)

Ronald Nikel (renikel@ix.netcom.com)
Sat, 31 Jan 1998 10:25:51 -0800

Lawrence,

I would step back and ask myself a set of questions :

1) how long are the lines
2) what is the fastest edge rate of the signal

Having asked myself those questions if your transmission line
is about the same length (in time) as the fastest edge then the answer
is : don't worry about the capacitors leave them off, except for charge
storage for the power supply.

Having said that lets make the assumption that your transmission
lines are transmission lines. The answer to your question is
swap V5 and GND in your plane pair and reduce the dielectric
thickness to 5 mil (less would be better, but I am sure
that you want high volume production). Your plane capacitance will
be around 180pf/in^2 so you should have adequate capacitance
built in, but if you are concerned I would use 0603 surface mount
caps with a value of 100-500pF in areas of high routing density
between power and and ground. Since your plane inductance will
be rather low your plane capacitance should be rather good and have
a high resonant frequency and be able to supply any dynamic current
required by fast edges.

Oh I forgot to explain why you should swap V5 and gnd. Your stackup
is not taking full advantage of power to ground coupling because
you have taken your two power planes and used them to sandwitch
your dual stripline layer. this could possibly cause problems with
ground path returns if you do not have adequate coupling in certain
regions.

Ron.

----------------------------------------------------------------------------
-----------------------
Ronald Nikel
Circuit Interconnect Technology Consulting
TriCN Associates
PH/FX : (415) 824-4154
CELL/PGR/VMAIL : (415) 509-1283