Mathematica 9 is now available
Student Support Forum
-----
Student Support Forum: 'NDSolve with very small step sizes' topicStudent Support Forum > General > Archives > "NDSolve with very small step sizes"

Next Comment >Help | Reply To Topic
Author Comment/Response
Jack Kohoutek
07/18/08 2:24pm

I am trying to use NDSolve to solve a diffeq with very small step sizes. Whenever I add a certain term, I get the error:

step size is effectively zero; singularity or stiff system suspected.

I think it thinks the step size is too small because I am using very small numbers (~10^-40). When I convert all the units in the problem from meters to nanometers it works, because everything is multiplied by 10^9 (lengths) or 10^18 (areas). I feel like there should be a way to tell it that the small step size is very small but not effectively zero. I think that the commands I would use would be MaxStepSize or MaxStepFraction but I don't know the default values. I am already using MaxSteps->10^6 and I don't think it can go any higher.

I would really like to be able to use my computer to its fullest, doing calculations with very small step sizes or large amounts of steps, but right now it seems that I can't.

I have attached a notebook with the two different pieces of code, the one that works and the one that doesn't. Maybe someone can help me get the first one to work. The ESFamp/(d-xyz)^2 term is the one that gives the problem.

Thanks for reading the long post and thanks in advance for any help.

Jack

Attachment: test.nb, URL: ,

Subject (listing for 'NDSolve with very small step sizes')
Author Date Posted
NDSolve with very small step sizes Jack Kohoutek 07/18/08 2:24pm
Re: NDSolve with very small step sizes yehuda ben-s... 07/19/08 4:28pm
Re: Re: NDSolve with very small step sizes Jack Kohoutek 07/20/08 10:20am
Re: Re: NDSolve with very small step sizes Jack Kohoutek 07/20/08 10:34am
Next Comment >Help | Reply To Topic