Re: rotating constraints

From: Anna Modzelewska (
Date: Wed Sep 21 2005 - 04:22:30 CDT

I would like to calculate forces acting on atoms in a simulation with rotating constraints as well as the torque.
I found the mail (cited below) with some instructions on the mailing-list but I still don't know how to do it in the practice.
Does anyone have scripts that extract the radius vector from dcd file and calculate forces?
Maybe someone could give me more detailed technical instructions how to do it?

Thanks in advance
  ----- Original Message -----
  From: Jim Phillips
  Sent: Monday, February 02, 2004 9:27 PM
  Subject: Re: namd-l: rotating constraints

  ---------- Forwarded message ----------
  Date: Mon, 2 Feb 2004 12:45:13 -0600 (CST)
  From: Aleksei Aksimentiev

  Hi Himanshu,

  In order to compute the torque you need to record actual positions of the
  atoms to which you apply the spring forces. This should be done after
  your namd run is complete. You take the dcd trajectory, and extract the
  radius-vector of all atoms subject to the force. Next, you need to compute
  force. For that, you use the 2K(R-Rt(t)) formula, where R is the actual
  position of the atom, Rt(t) is the position of the rotating template.
  Ones you know all forces and radius vectors, you can compute the torque
  as T=[RxF].


> hi all -
> referring to the rotating constraints feature in namd. We know
> that the force is 2K(R-X). Now R is the current reference position
> which should hence lie in the rotational circle in which we rotated the
> atom - and as we've already given the angular velocity of rotation we
> already know the current reference position at any time.
> My problem is about X. Now though the manual doesn't say it - X should be
> the current position of the atom(i'm sorry if i'm wrong and ofcourse all
> these quantities are vectors).
> My problem is in calculating the torque which i've put on each atom.
> Though it appears easy to do - how can the torque be zero if i have put my
> spring constants so stiff that the atoms being rotated follow their
> reference point very closely.
> Now - i am sure that i am wrong somewhere - could you please point out
> where i am wrong - and hence how the torque applied on the atom
> calculated.
> thanks
> himanshu
> iit delhi

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:41:09 CST