Global precision and accuracy.

*To*: mathgroup at smc.vnet.net*Subject*: [mg22161] Global precision and accuracy.*From*: iioannou at u.washington.edu (I. Ioannou)*Date*: Thu, 17 Feb 2000 01:23:49 -0500 (EST)*Organization*: University of Washington, Seattle*Sender*: owner-wri-mathgroup at wolfram.com

Hi all, I have been trying to adapt someone elses notebook to my purposes. What the notebook does is to solve a polynomial equation of varying order, to produce dispersion relations. (plasma waves, to be precise) All works fine when I plot the graphs. However, I need to look at the independent variable from 1x10^-7 to 4x10^-7, that is, I need to figure out the dispersion relations for that band of frequencies. (this corresponds to 10-40 kHz) Anyway, it seems that I encounter numerical problems. The notebook seems to behave from 10^-6 and above. I assume I need to specify better precision and accuracy, globally. I have tried to do this locally also, with no success. The main problem is in a FindRoot call, as well as in the diagonalization calls. If I could tell Mathematica to manipulate and return the nos with 10 digits throughout, I bet everything would work. (barring an error in the routines I was supplied) I've tried with FindRoot WorkingPrecision and such, but nothing good came out. If there is any suggestion as to how to force Mathematica to increase precision and accuracy globally, that would help me immensely. Or else I will have to modify the equations one by one and rewrite the whole thing. Any suggestions much appreciated. Thanks, John -- Ioannis I Ioannou phone: (206)-543-1372 g-2 group, Atomic Physics fax: (206)-685-0635 Department of Physics University of Washington e-mail: iioannou at u.washington.edu