[gmx-users] force constant

[Gavin Melaugh]

Hi all

I am running umbrella sampling to generate a potential of mean force
curve. Most of the histogram generated from g_wham are fine, and some
are not. I have focused my attention on one of these histogram and
played about with a few parameters so that I can get tobe the correct
shape. However using a force constant of 5000 give me a much sharper
distribution than when using a force constant of 10000 (which is broader).
Why is this so? The input files are the exact same apart from the
pull_k1 value. The ref distance is also the exact same 2.59 nm. I double
checked the fluctuations of the distances between the centres of mass of
the two molecules using g_dist, and again the higher force constant
gives greater fluctuations. How can this be?


Many thanks in advance

Gavin