Gaussian98W and charged molecules (again)



Hi --
 I'm sorry to bother everyone again...  But I have another question about
 Gaussian...
 I was finally able to get Gaussian to geometrically optimize my charged
 (-1) molecule with the following Route specification:
 # PM3 SCF=vshift=500 SCF=(MaxCyc=1200) Opt
 However, when I ran the job again and added the Freq keyword, although
 the geometry optimized, the SCF energy calculation in the Freq part of
 the job oscillates again until the job terminates with a
 UNABLE TO ACHIEVE SELF-CONSISTENCE error.  Why does this happen?  If I
 have the following route specification
 #PM3 SCF=vshift=500 SCG=(Maxcyc=1200) Opt Freq Test
 Does Gaussian know to use my prior specifications for finding the SCF
 (i.e. the Vshift and the increased MaxCyc) or does it use it but still
 fail?
 Does anyone have any suggestions?
 And thanks again for all your previous suggestions!
 Connie