# [gmx-users] switch potential function gromacs 3.3.2

Claus Valka lastexile7gr at yahoo.de
Wed Jul 23 15:54:57 CEST 2008

```Dear Sir or Madam,

For the version 3.3.2 the gromacs switch potential function is the following :

Elja1  = (4.d0*eps1*sig1**6) *                            &
-(1.d0/r(i)**6)     *                            &
(1.d0 - 10.d0 * (r(i)-r1)**3 * 1.d0/(rc1-r1)**3 &
+ 15.d0 * (r(i)-r1)**4 * 1.d0/(rc1-r1)**4 &
-  6.d0 * (r(i)-r1)**5 * 1.d0/(rc1-r1)**5)
Elja2  = (4.d0*eps1*sig1**12) *                           &
(1.d0/r(i)**12)      *                           &
(1.d0 - 10.d0 * (r(i)-r1)**3 * 1.d0/(rc1-r1)**3  &
+ 15.d0 * (r(i)-r1)**4 * 1.d0/(rc1-r1)**4  &
-  6.d0 * (r(i)-r1)**5 * 1.d0/(rc1-r1)**5)

Switch = Elja1 + Elja2

Remarks:
i)   This is tested only when in the mdp file the vdw-type = Switch
ii)  The Elja1 is the dispersion part of the potential (not the force).
iii) The Elja2 it the repulsion part of the potential (not the force).
iv)  eps is the value of epsilon and sig is the value of sigma
v)   r1 =  rvdw-switch and rc1 = rvdw
vii) The function of the manual that sais that it applies Phi(r) both on switch and shift option doesn't seem to apply in the switch case.
vii) Someone can find out if someone searches at : a)/root/Desktop/software/gromacs/gromacs-3.3.2/src/mdlib/
b)The file is called tables.c
c)look under if (bSwitch), where it sais swi
d)for pow look at the beginning
e)for Vtab under the cases etabLJ6 and etabLJ12
f)for whole under if ((r_r1) && bSwitch) , that is the multiplication of the dispersion and repulsion part with swi
g)in order to have units of energy someone must to multiply the above with 4*epsilon*sigma**6 for the dispersion part and 4*epsilon*sigma**12 for the repulsion part

Yours Sincerely,
Nikos

__________________________________________________________