On the Source Implementation for the Leapfrog ADI-FDTD Method

Different source implementations for the leapfrog ADI-FDTD method including both current and hard sources are investigated in this letter. It is shown that, different from the conventional ADI-FDTD method, the leapfrog one always possesses asymmetry errors, while the hard source results in larger asymmetry errors than the current source. An optimal current source derived from the conventional ADI-FDTD method is then presented. In addition, it is found that the hard source is less accurate than the current source as the CFL number increases.