From: Jérôme Hénin (
Date: Thu Feb 03 2011 - 11:13:22 CST

Hi Alexandra,

Technically it would be possible to use a combination of dihedrals.
The main problem, however, is that periodic variables get a special
treatment. A combination of several periodic components may or may not
itself be periodic. Then it gets really difficult to handle things
properly in the code - so we decided to forbid this. I understand,
however, that this might be too restrictive in some special cases.
Note that what you are doing is risky: for example, an angle moving
from 165 to 175 degrees will increase the average a little - however,
a further 10 degree rotation will cause a large drop in the average.
That is only avoidable if each angle evolves within a somewhat
restricted range (in that case, the wrapAround option can be used to
ensure that the value will not cross a periodic boundary).

I am going to replace the error you got with a strong warning message.
If you confirm that you need this, contact me off-list to get the
modified code.


On 3 February 2011 11:30, Alexandra Nemkevich
<> wrote:
> I am  trying to set up a colvar that is an average of six dihedral angles.
> Although NAMD 2.7 User Guide says it's possible, I keep getting this error:
>  Error: superposition of periodic components is unsupported.
> Any suggestions?

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:19:46 CST