[gmx-users] MD run, polarizable molecules, EM did not converge

[Andreas Kring]

Hello all.

I've been searching the archives with no luck so here goes:

I am using shell molecular dynamics to simulate a hydroxide ion 
dissolved in water. I have been using the steepest descent minimizer 
(tried in the range from 300000 to 700000 steps) to minimize the system, 
but I'm not capable of getting Fmax lower than

Potential Energy  = -2.39282728357508e+04
Maximum force     =  2.22634509434073e+03 on atom 3
Norm of force     =  5.74522076456533e+01

Running the shell MD simulation with the output from the above 
minimization generates a lot of "EM did not converge" (see below). As 
recommended in the GROMACS manual, I'm using the follow parameters in 
the mdp-file:

emtol = 1
niter = 20
fcstep = 0

step 1: EM did not converge in 20 iterations, RMS force 307.457
step 2: EM did not converge in 20 iterations, RMS force 306.841
step 3: EM did not converge in 20 iterations, RMS force 308.809
...

Any ideas on how I can fix this?

Regards
Andreas