Linear dependence in geometry optimization

[Dave Moore]

Dear all,

I am having a problem with a fairly simple geometry optimization (mp2/VDZ
for dimethyl ether) in MOLPRO 2002.3 with a recently updated patch-set (I
believe it to be completely current).  No matter what symmetry I specify
(C2V, Cs, or C1), this optimization always crashed with an error message
like:

 ?ERROR IN ORTHS: LINEAR DEPENDENCY DETECTED FOR VECTOR  37.1
 OVERLAP= -0.4437E-15

after 1 or 2 steps.  I have gone through the input and output several times
looking for evidence of a careless mistake on my part, and I can't seem to
find it.  Is this a problem with MOLPRO or am I just overlooking something
silly?  I would appreciate any insights ... the input file is appended to
the end of this email.  Thanks in advance for any help.

Dave Moore
FOM Rijnhuizen
Edisonbaan 14
3439 MN, Nieuwegein
The Netherlands
+31+30 609-6826
FAX:   603-1204



****** BEGIN MOLPRO INPUT ******
***
mem,60,M
basis,vdz
gthresh, energy=1.0e-10, orbital=1.0e-10, optgrad=4.0e-5
gexpec, dm, sm, qm

file,2,methyl_ether_monomer.C2v_mp2.vdz.wf

geometry = {ANGSTROM;
O1
X1  O1 1.0
X2  O1 1.0   X1  90.0
C1  O1  rco  X1  acox  X2  90.0
C2  O1  rco  X1  acox  X2 270.0
H1  C1 rch1  O1 ahco1  C2 180.0
H2  C2 rch1  O1 ahco1  C1 180.0
H3  C1 rch2  O1 ahco2  C2  60.0
H4  C1 rch2  O1 ahco2  C2 300.0
H5  C2 rch2  O1 ahco2  C1 300.0
H6  C2 rch2  O1 ahco2  C1  60.0
}

 rco = 1.41004
rch1 = 1.09981
rch2 = 1.10938


 acox  = 55.285
ahco1 = 107.534
ahco2 = 111.783

hf;orbitals,2101.2
mp2; dm,8101.2; natorb,,,core=1
optg,energy=1.0e-8,grad=1.0e-5;

put,molden,methyl_ether_monomer.C2v.mp2.vdz.molden.dat;orb,8101.2

---
******   END MOLPRO INPUT ******