electromagnetic simulation using the fdtd method by dennis m sullivan

[mjl23]

Hi i was wondering if anybody had worked through any of the problems in this as i am currently stuck on chapter 2 i can not get any of the programs running and am getting a bit lost with what is needed in order to complete the programs, if anybody has any advice or help much appreciated..

 

[iyami]

Nobody I know got Sullivan's programs running without extra work. I gave up on the code somewhere in the first chapters and started writing my own code.
There is actually a Japanese textbook (Uno) which is quite reliable (Fortran code and if I remember correctly only one easily corrected error in the PML code).

As for English textbooks I think that Taflove's code is ok, but supposedly (I have not read it) Elsherbeni is easier to use.

Looking back at my own road to FDTD I would suggest you first write the programs in a high-level language
(something like python) you feel comfortable with. Later you can switch to C/Fortran or whatever 'fast' language you like.

Copying from a book is faster at first but if you want to become semi-competent at FDTD I would suggest that once you
understand the basic update equations you try to write the code yourself.

You will learn things like
- how do you deal with boundary conditions at the corners (which is rarely properly explained)
- understand how to use permutation of variables to intelligently transform the Ex update to Ey and Ez
- debugging (it is very important to learn that while your algorithms are simpler)

A last word of advice: do not worry about the slow progress at first if you try to do it the hard way.
I believe there are enough people here which could (and would be willing to) help you if your own code
does not work once you post the 'problematic parts' here

 

[mjl23]

Thank you for your reply i have actually made significant progress with this method now, i used the book only as refrence and started to work using matlab due to it's good visualisation tools and my general understanding of matlab was much better.
I also agree working at generating your own code is a good idea as my general understanding of this method has improved greatly, and a bit of algebra manipulation always helps me get into it aswell. my programs are comming along nicely.
if i need any help in the future i will be sure to ask
Many thanks for your reply

 

[iyami]

if i need any help in the future i will be sure to ask

I tend to check for posts in irregular intervals (usually at least once a week but sometimes only once a month). I will try to reply but you might have to wait for a bit.

Many thanks for your reply

You are welcome.