MRCI problems

[Nathan DeYonker]

I'm running into a small difficulties when running MRCI calculations. 

I'm using Molpro 2002.3 at Pacific Northwest National Lab.

When running an MRCI calculation, I get the following error:

 TOO MANY SPIN COUPLINGS:  1638 MXSPC=  1001

This number lessens as I increase the coefficient threshold for passing
mcscf configurations into the MRCI code, however, I have increased the
threshold to the point where only the reference configuration is being
passed to MRCI, thus rendering the point of using MRCI useless. The second
and third most important configurations have a coefficient of .125, so I
would like to include these in my MRCI treatment. Is my active space of 12
active MO's and 54 virtuals too large? Is there anything in my input that
looks suspicious?

the execution part looks like this:

fn=2.0619333345,nc=1.1875349326,lin=180.0
geometry={angstrom;
          fe;
          c,fe,fn;
          n,c,nc,fe,lin}

{wachters basis set removed for readability}

rhf;occ,13,4,4,1;closed,11,3,3,0;wf,39,1,5;shift,-.5

gprint,civector;

multi;core,5,1,1,0;closed,9,2,2,0;save,ref=6500.2;
wf,39,1,5;
noextra;
expec2,lzz;

ci;core,5,1,1,0;closed,9,2,2,0;select,6500.2,,0.13;wf,39,1,5;
print,orbitals,ref=1;

Thank you for any help!
Nate

 --------------------------------------------------------------------
  Nathan DeYonker              | phone: 706 542-2067
  Ctr. for Comp. Quantum Chem. | fax:   706 542-0406
  University of Georgia        | e-mail: nate at ccqc.uga.edu
  Athens, GA  30602-2556       | http://zopyros.ccqc.uga.edu
 --------------------------------------------------------------------