Re: [ESPResSo-users] Tabulated angle potential

[Vincent Ustach]

Hi Tristan,

Thank for your response. When I use the tabulated 2-body bond potential I do include the scaling factor and it works just fine. I am having trouble with the 3-body angle potential. Last month I discussed with Axel Arnold on the mailing list that it might be OK to leave off the 1/r scaling factor. If I did include the 1/r, would you suggest I use the equilibrium bond length? The length to use for tabulated angles is not clear to me.

Best Regards,

Vincent

--Vincent Ustach
  University of California, Davis

On Sun, Aug 24, 2014 at 11:41 PM, Tristan Bereau <address@hidden> wrote:

Hi Vincent,

It's hard to tell what exactly went wrong here, but just to point out
the obvious: are you aware that Espresso has a rather special way of
formatting the tabulated potential file? Specifically, the force is
rescaled by the distance:

F(r) = -V'(r)/r

where V'(r) is the derivative of the potential.

Best,
Tristan

On Mon, Aug 25, 2014 at 9:09 AM, Vincent Ustach <address@hidden> wrote:
> Hello everyone,
>
> I am working on simulating a polymer with tabulated potentials. I noticed
> that the distribution of the bond angles is centered about the wrong angle,
> so I checked my method by comparing the built-in angle potentials-
> angle_cosine, angle_cossquared, and angle_harmonic with equivalent tabulated
> potentials. The three attached .m files creates data that I've copied into
> files for Espresso to read- all I do is add the header
>
> #1001 0 3.14159
>
> Attached please find the script I have used to run the simulation as well as
> histograms for the built-in and tabulated potentials. Each set of data
> represent bond angle distributions for a 100-mer (entered as an argument of
> the execution command).
>
> Any advice to input the tabulated potential correctly is most welcome. And
> sorry to attach so many files!
>
> Thank you,
>
> --Vincent Ustach
>