[Wien] total energy with SO+orb(H_ext)

[mazin]

I've been testing the orbital potential module in the 3rd mode (external 
field).
The case.inorb file is as follows:
            3           1           1
PRATT 0.2
   1  1  2 (for Mo, or  1  3  0  1  2  for Cu)
   1000
   0.0000000E+00  0.0000000E+00   1.000000

I am varying the field and plotting the total energy as a function of 
the induced, comparing it with the fixed spin moment calculations. My 
testing materials are Cu and Mo, so I do not expect a considerable 
difference. If I run calculations without spin orbit, i.e., runsp -orb, 
I get a perfect agreement. If I add so, that is, runsp -orb -sp, I get 
an unphysical result, namely the total energy decreases with the field.

However, if I plot E+L*Hext instead, where L is the calculated orbital 
moment, I recover a decent agreement with the FSM calculations.
Of course, one can define the total energy including or excluding 
interaction with the external field -L*H, but there are several reasons 
not to include it here:
(1) it is not included for spin moments
(2) by not including it, one can compare directly with the FSM calculations
(3) If it is not included, one can calculate the full susceptibility by 
differentiating the total energy w.r.t. the external field.