[gmx-developers] COM pulling with pull_geometry = direction_periodic

[Magdalena von Essen]

Dear all,
I am posting to the developers list because this question was not 
answered by post in the users list.

I am using "pull_geometry=direction_periodic" for constant force pulling 
(pulling force in z-direction = 50pN) of a protein molecule. With this 
setting, the pulling force is considerably smaller compared to 
"pull_geometry=direction" option (pulling force in z-direction = 300pN). 
The protein unfolding trajectory is comparable for both settings. The 
pull_geometry is the only difference in my .mdp file, all other 
parameters and values are according to gromacs manual.

What is the reason for such significant deviation?
Is it simply that the pulling force is not added to the virial energy?
Does the removed z-axis pressure control have large effects? (on the 
pulling force, on the viscosity of the system, etc.)
Does the size of the system (N particles) have large impact on the force 
magnitude in both of the pull_geometry settings?

Thanks for any explanation.

Greetings,
Magdalena