KRCI with Olsen correction activated - convergence problem

[Adam Koza]

Dear Dirac Experts,

I have been trying to run KRCI calculations for a kind of exotic molecule, YbCr  (six unpaired electrons in the ground state), for some time. I designed GAS that I think is more or less reasonable:

4                                                                                                   
3  6  / 3       # p elec of Yb                                                                                     
19 20 / 7     # f elec of Yb                                                                                      
26 28 / 7     # d and s elec of Cr + s elec of Yb                                                                                      
28 28 / 179      virtual

However, I encountered a problem with CI iterations for particular excitation schemes from given GAS: whenever I excited more than 2 elec from p shell of Yb: “ 3  6 / 3” or more than one from f shell of Yb: “18 20 / 7,”  at some point, “Olsen correction” (an example of such a behaviour in the attached output) is activated, which completely destroys the iteration procedure (I think so). Is there a way to turn it off? 

For example, for this excitation pattern, there is no problem (maybe only quite large number of iterations required to obtain convergence):

4                                                                                                   
4  6  / 3       # p elec of Yb , 2holes                                                                                   
19 20 / 7     # f elec of Yb  , 1hole                                                                                    
26 28 / 7     # d and s elec of Cr + s elec of Yb    , 2 holes                                                                                  
28 28 / 179      virtual
 

In the attached example, I cancelled the job, but from previous experience, I know that energy is going down absolutely too much.

best regards,
Adam