Mopac93: Structure of excited states

 I am interested in the relative energies and structures of the first excited
 states (triplet and singlet) of a series of compounds.  I have obtained the
 energies relative to the ground state using the keywords: MECI C.I.=2.
 I now want to refine the structures of these states, but after 10 cpu-hr on an
 Indigo2, the gradients are still very large (> 20) and appear to be
 I started the calculation with the optimized ground state structure, and used
 the keywords: AM1 EF MECI C.I.=2 ROOT=2 (for the triplet, ROOT=3 for the
 singlet).  I have also done the calculation adding SHIFT=15, in an attempt to
 damp the oscillation.
 This is my first attempt at working with excited state structures. . .is this
 typical behavior, and I just need to be patient (sigh!), or have I missed
 If there is interest, I will summarize useful replies, for which I thank the
 authors in advance.
   Stephen B. Bowlus, Ph.D.                Computer-Aided Molecular Design
                                           Research Division
   e-mail: bowlus &$at$&               Sandoz Agro, Inc.
   Phone:  + 1 415 354 3904                975 California Ave.
   Fax:    + 1 415 857 1125                Palo Alto, CA 94304