bird 28 comments

Hello Jon, and others:

     While I think that almost all realistic packages do have intersperced
ground I have to agree with Jon -- there can be a situation of signal, GND,
signal coupling and thus the restriction becomes unworkable. (There is also
Jon's original example of no grounds and the leads couple to there nearest
neighboor. Even if this is unlikely the spec should handle this case.)
I guess my question now is what simulation tools will do if actual section
lenghts are included. The scenario as I see it is:

 1. The field solver assumes a bunch of parallel wires of infinite length,
     and reports coupling 'per unit length'.

 2. The coupling per unit length, and the actual length of each section
     is entered into the package file.

 3. Now, can existing simulation tools handle the actual, unequal lengths?
     For example, if one is stimulating a section of sig2, and it couples
     to sig1 and sig3, would a simulation tool assume that all three sections
     are of equal length (perhaps use the average length). Does a tool have to
     include the extra length of the longer sections as an uncoupled bit of L/C?
     

    This is the question I think we need to discuss Friday. Anyway, thanks for
your comments.

          Regards,
          Stephen Peters
          Intel Corp.
     
Received on Wed Jul 19 16:30:54 1995