The SHORT answer is NO, you cannot change the width of the ground to equal
the width of the trace and keep your 50 ohms!
The slightly longer answer is,
If you look and the field lines (lines of flux?) around a trace in free air,
they will be circular around the trace, but more dense close to the trace
and less strong the further away they get. It is a square of the radius I
think.
Now, if you move a plane close to the trace you will see the lines snap
towards the plane, and they will be dense right between the trace and the
plane, but moving out from the edges there will be lines curving gently
towards the plane. I can't draw this with ASCII characters, sorry, but if
you look at a cross section with a plane on either side it will have a sort
of hourglass shape, ok?
So now if you cut the planes down to the width of the trace most of them
will try to get to the plane, but many will be fringing around the sides,
and you won't have your 50 ohms anymore.
What we used as a general rule (we have done the same thing you are, kind of
making a coax effect on board layers) is to make the ground width FIVE TIMES
the trace width. Even so this will not be ideal, but past that the effect is
so minimal you probably wouldn't care.
onward thru the fog...
Jack
p.s. unfortunately, to calculate at what point the trace is 10% or 20% of
ideal takes a fairly interesting calculation, something like delta2V=0, and
you keep calculating all different points of potential in the field until
they all add up to zero or something like that, I just went with the 5X rule
and it worked great. Good Luck
-----Original Message-----
From: Doug [mailto:[log in to unmask]]
<mailto:[mailto:[log in to unmask]]>
Sent: Thursday, April 06, 2000 10:38 AM
Subject: Long Question Needs Short Answer ...
Suppose there's a trace construction as follows.
It's 50 ohm controlled impedance with a separation
"d" and a width of 5 mils between two ground planes
in a symmetrical construction. I'm assuming the
controlled impedance is calculated by assuming the
ground planes go off to infinity.
CASE #1:
ground plane
---------------------------
5 mils d
-----
d
---------------------------
ground plane
Now, for some reason, someone wants to start cutting
back the ground planes to a width W without changing
separation "d" of the trace to the ground planes.
Shown as below.
CASE #2:
ground plane
| <---- W ----> |
-----------------
5 mils d
-----
d
-----------------
| <---- W ----> |
ground plane
Question: Ignore fab tolerances for the moment.
At what point does the controlled impedance of the
trace begin to be compromised by 1%, 10%, 20% ... ?
Could I in fact reduce the width of the ground planes
down to 5 mils, equal to the width of the trace,
without compromising impedance?
Regards, Doug McKean
##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in
the body:
To subscribe: SUBSCRIBE TECHNET <your full name>
To unsubscribe: SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Keach Sasamori at [log in to unmask] or
847-509-9700 ext.5315
##############################################################
|
