Jack Olson wrote:

<snipped material> 
>  Many of the RF designs here were done in AutoCAD, and the only reason I
>  have been given for this was that it was easier to chamfer the corners
>  of the delay lines (fairly wide 50 ohm traces on 20mil teflon). I may be
>  wrongly assuming that the reason for the chamfer was to minimize the
>  effect of the impedance change "around the corner", but my new CAD
>  package can do curved traces which suggests NO mismatch around bends...
>  constant width everywhere, right?
>  Unfortunately, no one here has any data to support the need for
>  chamfers. (I have to painstakingly enter precise polygons if chamfers
>  are really required, so I would rather avoid it).
>  We are working with 1.9GHz, and are moving to Rogers 4003 material.
> 
>  Can anyone with RF experience comment on this?

Hi Jack, 

The issue of corners comes up every now and then. 
Part of the issue is that as frequency goes up, 
the turn of the trace is part of a turn for a loop. 
i.e. inductor. 

The other part I have always used I will try to 
address in the following. 

A side trace-end-on point of view, current 
distribution in the ground plane will look 
like the following... 

                         ______
               Trace    |      |
                        |______| 
                        ........
                       .        .
                     .            .
                   .                .
                .                      .
Ground Plane .____________________________.

Strongest underneath, dropping off exponentially 
as you move away from the trace to some other point 
in the ground plane. Put two traces very close to 
one another and their respective current distributions 
in the ground plane will conflict with one another. 

Now, let's add a right angle turn to your trace 
and you get more of a mess.  The current distribution 
on the outside bend drops off much more rapidly than 
the above. Why? It's spread across a 270 degree angle. 
The current on the inside of the bend does not drop 
off so rapidly. Why?  It's compressed into a 90 degree 
angle.  This spreading and compression causes an 
"impedence bump". The current is not evenly distributed 
as before. The only way this can be explained is if there 
is a discontinuity.  In this case, it's a discontinuity 
in the current that "appears" as a discontinuity in 
impedence. Put another trace close to your first and 
it just adds to the mess. 

90 degree corners modified to two 45 degree 
corners or a corner with a radius helps. 
And keep your traces fairly spaced from one another. 

Hope this helped...

***************************************************************************
*   The mail list is provided as a service by IPC using SmartList v3.05   *
***************************************************************************
* To unsubscribe from this list at any time, send a message to:           *
* [log in to unmask] with <subject: unsubscribe> and no text.*
***************************************************************************
* If you are having a problem with the DesignerCouncil, please contact    *
* Dmitriy Sklyar at 847-509-9700 ext. 311 or email at [log in to unmask]      *
***************************************************************************