G90/MOPAC: Problem solved -- new problem appeared

[milan "-at-" helix.nih.gov (Milan Hodoscek)]

 Dear netters,
 Thanks to Ross Nobes the problem with differences in energy reported
 by Gaussian MNDO method and MOPAC is solved, but grep on mopac
 directory reports the following:
 grep 0\.52 *.f
 analyt.f:      A0=0.529167D0
 block.f:      DATA QQPM3 (  7)/       0.5293383D0/
 block.f:      DATA ADPM3 (  8)/       0.5299630D0/
 block.f:      DATA AQPM3 ( 34)/       0.5283737D0/
 block.f:      DATA GUESP3( 52,2)/       0.5242430D0/
 delri.f:      A0=0.529167D0
 ders.f:      A0=0.529167D0
 diat.f:      R=R1/0.529167D0
 diat2.f:      RAB=R12/0.529167D0
 esp.f:      BOHR = 0.529167D00
 esp.f:      DATA BOHR/0.529167D0/
 esp.f:      BOHR = 0.529167D00
 esp.f:      DATA BOHR/0.529167D0/
 esp.f:      DATA BOHR/0.529167D0/
 esp.f:      DATA BOHR/0.529167D0/
 gover.f:      R=R/0.529167D0
 hcore.f:                     HTERME = -0.529177D00*DD(NI)*EFIELD(1)*FLDCON
 hcore.f:                     HTERME = -0.529177D00*DD(NI)*EFIELD(2)*FLDCON
 hcore.f:                     HTERME = -0.529177D00*DD(NI)*EFIELD(3)*FLDCON
 powsq.f:  350 BMAT(K,ID) = SIG(K)/0.529167D0
 repp.f:      DATA A0/0.529167D0/ ,EV/27.21D0/, EV1/13.605D0/, EV2/6.8025D0/,
 HTERME has different constant then the rest of program and this values
 go into electron core interaction matrix.
 Regards -- Milan Hodoscek