Questions about SCF convergence again with Gaussian98W and Hyperchem

[Connie Chang <cc236(-(at)-)cornell.edu>]

Hi everyone --
 thank you for all your good suggestions...
 I'm still having trouble getting an optimization and frequency
 calculation to work for a charged molecule with 120 atoms.  I've tried
 with both Hyperchem and Gaussian98W using the PM3 method.  And I've
 tried some fo the suggestions given by listmembers such as increasing
 the maximum number of iterations to 500, setting scf=sleazy, setting
 scf=qc, etc...  I still get the UNABLE TO ACHIEVE SELF CONSISTENCE error
 after (at most) one optimization step.  So the energy will converge
 once, it moves the atoms by a little bit for the optimization and the
 energy fails to converge.
 I have the same problem with hyperchem.  Hyperchem does the neutral case
 just fine, but as soon as I add a charge of -1 and a multiplicity of 2,
 the energy does not even get close to something reasonable (it's large
 and positive).
 Does anyone have any further suggestions on what I can do to get either
 of the programs to work?  Hyperchem does work for a smaller molecule
 (~70 atoms), but Gaussian does not.  Does the size of the molecule make
 it more difficult for these programs and methods to find the minimum
 energy?  I tried a charged water molecule in Gaussian and it worked
 fine....
 thanks.
 Connie