Re: [ESPResSo-users] Fwd: Scaling of tabulated angular potentials

[Hender Lopez]

Hi Paul,

I remember having the same question quite a time ago and after getting
the comments from the list and looking at the code I found that the
the straight answer to you question is:
For an angle tabulated potential the file containing the potential has
to have the following structure:
- First line: start with # and it should have 3 elements, 1) the
number of points of the tabulated potential, 2) min angle (in rad) and
3) max angle (in rad).
- Then you should have 3 columns: 1) angle (lets call it theta), 2)
dU(theta)/dtheta and 3) U(theta) where U(theta) is your potential that
depends on theta.

Now I don't remember exactly which files I checked but I do remember
that my sims behaved well. One way to test it, is by creating a
tabulated potential for one of the bond-angle functional form
implemented in espresso and run a simple system and compare results
(tabulated vs functional form implemented in espresso). For example
try two beads join by an bond-angle interaction and get the energy for
different angles.

Best,

Hender


On Wed, Jul 22, 2015 at 1:26 PM, Paul Peterson
<address@hidden> wrote:
> ...does that mean that I SHOULD NOT divide by anything?
>
> Put another way: should I just calculate the force as the derivative (with
> respect to theta) of the angular potential, without scaling factor, UNLIKE
> what I do with linear potentials?
>
> Maybe this is the solution. In this way, the potential would be purely
> angular, that is, already independent of distance.
> Any feedback?
>
> Thanks in advance,
>
> Paul
>
> ---------- Forwarded message ----------
> From: Paul Peterson <address@hidden>
> Date: 2015-07-22 11:38 GMT+02:00
> Subject: Re: [ESPResSo-users] Scaling of tabulated angular potentials
> To: Rudolf Weeber <address@hidden>
> Cc: ESPResSo Users <address@hidden>
>
>
> Hi Rudolf,
>
> thanks for your reply.
> Your reasoning makes perfect sense, but the point is: when I define the
> tabulated angle potential I don't know what the distance (radius) will be.
> So what 1/r should I scale by?
>
> Paul
>
> 2015-07-22 11:31 GMT+02:00 Rudolf Weeber <address@hidden>:
>>
>> Hi Paul,
>> On Wed, Jul 22, 2015 at 11:07:24AM +0200, Paul Peterson wrote:
>> > I know this is not a new topic around here, but I've not been able to
>> > find
>> > an answer in these threads.
>> >
>> > I'm simulating a polymer with (among others) tabulated angular
>> > potentials.
>> > What is the 1/r scaling factor of force in this case? A distance doesn't
>> > seem to make much sense to me - distance is variable and depends on
>> > other
>> > potentials, of course.
>> > So what?
>> I think, the reasoning is that the energy U(theta) given by the potential
>> should equal the integral of force times distnace over the path, a particle
>> travles when the angle is increased from 0 to theta.
>> As the distance the particle travels increases when the radius is larger,
>> the force has to be smaller.
>>
>> Regards, Rudolf
>>
>
>