Loading [MathJax]/extensions/TeX/boldsymbol.js
\newcommand{\uex}{{u_{\small\mbox{e}}}}
\newcommand{\half}{\frac{1}{2}}
\newcommand{\tp}{\thinspace .}
\newcommand{\Oof}[1]{\mathcal{O}(#1)}
Solving for the numerical frequency
The scheme
with u^n=I\exp{(i\omega\tilde\Delta t\, n)} inserted gives
-I\exp{(i\tilde\omega t)}\frac{4}{\Delta t^2}\sin^2(\frac{\tilde\omega\Delta t}{2})
+ \omega^2 I\exp{(i\tilde\omega t)} = 0
which after dividing by I\exp{(i\tilde\omega t)} results in
\frac{4}{\Delta t^2}\sin^2(\frac{\tilde\omega\Delta t}{2}) = \omega^2
Solve for \tilde\omega :
\tilde\omega = \pm \frac{2}{\Delta t}\sin^{-1}\left(\frac{\omega\Delta t}{2}\right)
- Phase error because \tilde\omega \neq \omega .
- Note: dimensionless number p=\omega\Delta t is the key parameter
(i.e., no of time intervals per period is important, not \Delta t itself)
- But how good is the approximation \tilde\omega to \omega ?