Dan McMahill - 2004-09-27

I'm interested in using tnt-mmtl to help evaluate the accuracy of a transmission line tool I've been working on.  My tool is not a field solver but used some closed form approximations. 

My question is how one estimates the error due to a finite number of segments used for the conductor and dielectric.  Are there any guidlines for how many segments to use to keep any errors below a certain value?  The temptation is there to put in 1000 segments or some such ridiculously large value and just let the computer crank away for a few weeks generating curves for me.  However the manual said something which makes me wonder if one can go overboard and produce an effect which is the opposite of the desired one.

Thanks for any insight. 

I still need to dig out my IEEE Xplore password to see if I can get the references online.

Thanks
-Dan