Re: [ESPResSo-users] Sistematically decreasing of Temperature with DPD thermostat

[Christoph Junghans]

2014/1/15 Stefan Kesselheim <address@hidden>:
> Dear Salvador,
>
> On Jan 15, 2014, at 4:11 PM, Salvador H-V <address@hidden> wrote:
>
>> And now the temperature is fluctuating around 1.6$$... I had to use a very 
>> high value of K to keep the particle's center distance around 1.0.
>
> OK. This means you script is not terribly buggy :-).
>
>> So, I have to think more carefully what Stefan said: "Thermalization of 
>> rotational degrees of freedom is entirely different in Langevin than what 
>> happens when combining rigid bodies with DPD (this so far probably nobody 
>> has done)."
>
> Let me ask you briefly one thing: Did you use rigid body dynamics with 
> virtual sites or did you use the bond length constraint thing. The second 
> should be a RATTLE or SHAKE implementation. I'm not exactly sure how it works 
> and I'm not sure if it works or anyone has tried  it. Anyone?
It might be possible, that RATTLE soaks up parts of the DPD forces
alone the connection bonds. And if your DPD cutoff is so small that
the interactions between multiple dumbbells cannot compensate for that
this will lead to the wrong temperature of the system. You could test
that by increasing the DPD cutoff and/or excluding DPD interactions
between the dumbbells using inter_dpd instead (I don't remember if the
thermostat DPD respect exclusions.)

Christoph
>
> Cheers
> Stefan
>
>
>



-- 
Christoph Junghans
Web: http://www.compphys.de