NMR, ZORA, and SCF convergence


Click here for full thread
Just Got Here
I have a rather large paramagnetic molecule (65 atoms) that I would like to calculate NMR shielding values for. Here is my input file:
echo
start nmrepr
memory stack 500 heap 250 global 750 mb

charge -2
geometry noautoz units angstrom nocenter
  symmetry c1
  load geom.xyz
end

basis "ao basis" print
  * library Def2-SVP
end

relativistic
  zora on
  zora:cutoff_NMR 1d-8
  zora:cutoff 1d-30
end

dft
  odft
  mult 3
  xc pbe0
  iterations 50
  convergence lshift 0.75 damp 90 rabuck 30
  semidirect filesize 100000000
end

property
  gshift
  hyperfine
  shielding
end

task dft property

I seem to obtain SCF convergence in 28 iterations, giving the following final iteration:
   convergence    iter        energy       DeltaE   RMS-Dens  Diis-err    time
 ---------------- ----- ----------------- --------- --------- ---------  ------
 d= 0,ls=0.8,diis    28  -2341.0843647713 -9.40D-08  1.98D-06  3.46D-08  5265.9
                                                     5.38D-07  4.47D-08

After the ZORA property calculations, the CPHF module began and gave me an error:
SCF residual:    3.7395528743051777E-002
 ------------------------------------------------------------------------
 cphf_solve2:SCF residual greater than 1d-2                   0
 ------------------------------------------------------------------------

How is that the case after reaching SCF convergence earlier? And how do I try to lower the SCF residual in the CPHF module?