AW: Exploding vdW energies in automatically generated POPC membranes

From: Norman Geist (
Date: Tue May 07 2013 - 00:16:20 CDT

Hi Lucas,

if you see energies looking like 999999.9999, you likely have superimposed
atoms which the minimizer won't get rid of. Try to find the atoms causing
the problem. Remember the periodic boundary condition. If your box is too
small f.i., the wrapped atoms will superimpose at the box boundarys when the
simulation starst. Additionally as I think your membrane is periodic in X
and Y direction, and the ionize command, if you got ions, doesn't know that
maybe, check if you have ions beside the membrane, that will be treated as
wrapped into the membrane at simulation start, what would also cause
superimposed atoms.

Norman Geist.

> -----Ursprüngliche Nachricht-----
> Von: [] Im
> Auftrag von Lucas
> Gesendet: Montag, 6. Mai 2013 22:34
> An:
> Betreff: namd-l: Exploding vdW energies in automatically generated POPC
> membranes
> Dear all,
> '
> I was trying to reproduce the membrane tutorial with my own protein,
> but I kept running into the "Constraint failure in RATTLE algorithm"
> error while trying to minimize the full solvated system with only the
> lipid tails allowed to move. I have tried some suggestions I've found
> in the archives for those errors to no success - energies always go
> too high immediately, then at some point I get that error.
> Some posts I've read said that automatically generated membranes
> shouldn't result in clashes, so I kept looking elsewhere to see what
> could be wrong, but after doing a few calculations using NAMD Energy,
> it seems it really is all down to the membrane. If for example I run
> the membrane builder with POPC, 160x160 size and c27 topology, and
> then run NAMD Energy, I get this:
> Frame Time Bond Angle Dihed
> Impr Elec VdW Conf Nonbond
> Total
> 0 0 +28971.9 +59388.9 +24331.3
> +341.334 -34635 +1e+10 +113033 +9.99997e+09
> +1.00001e+10
> So it seems that even before I add anything else, van der Waals
> energies are already too high. What am I doing wrong?
> Best regards,
> Lucas

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:11 CST