Fix torque based constraints for negative moments.

The constraint magnetic field that is calculated based on the torque
is now muliplied by the sign of the magnetic moment in the local
frame of reference.
The torque has to be understood as an energy related to a change in
the direction of the local frame of reference. This can differ from
the direction of the magnetic moment by a sign. For the constraint
fields, the energy related to a change of the direction of the moment
is needed, so this sign has to be put at this point.
(This is probably somewhat confusing, I'll add a better explanation
and probably some sketch to the documentation of these features.)

source/KKRnano/source/bfield/torque.f90

@@ -164,7 +164,7 @@ contains
 
     double precision, dimension(3) :: dir, mag_mom_dir
     double precision, dimension(3) :: old_b_constr
-    double precision :: mag_mom_len
+    double precision :: mag_mom_len, mag_mom_sign
 
     ! If the current iteration is not in the window the magnetic fields should be
     ! applied, return without changing the potential
@@ -177,13 +177,16 @@ contains
     dir(2) = sin(theta) * sin(phi)
     dir(3) = cos(theta)
 
-    ! Calculate direction and absolute value of the magnetic moment
+    ! Calculate direction and absolute value of the magnetic moment and the
+    ! sign the magnetic moment has in the local frame of reference
     mag_mom_len = sqrt(dot_product(bfield%mag_mom(:), bfield%mag_mom(:)))
     mag_mom_dir = bfield%mag_mom(:) / mag_mom_len
+    mag_mom_sign = sign(1., dot_product(mag_mom_dir, dir))
 
     if (constr_mode == 3) then
       bfield%bfield_constr(:) = bfield%bfield_constr(:) - &
-                    (bfield%mag_torque(:) / mag_mom_len) * constr_bfield_mixing
+              ( mag_mom_sign * (bfield%mag_torque(:) / mag_mom_len) * &
+                constr_bfield_mixing )
     else if (constr_mode == 2) then
       old_b_constr = bfield%bfield_constr(:)
       bfield%bfield_constr(:) = old_b_constr - dot_product(old_b_constr,dir)*dir - &