Re: [ESPResSo-devel] error in dawaanr interaction

[Juan Jose Cerda]

Dear Markus,

     L_b is in magnetostatic interactions in Espresso as a "dummy" constant by 
analogy with the electrostatic interactions. If you use  T=1, nothing to worry 
about, set it to 1 and end of the history. When you must be careful is when you 
set T to a different temperature, then,  to do things correctly you must set 
the magnetic l_b = 1/T to work with a dipolar interaction which is independent 
of the value of T. The reason to do that is  due to how Espresso treats l_b cte 
and multiplies it by T, when computing forces,etc . I know is wierd as hell, 
but at the point we were programming the original subroutines for 
magnetostatics we found better to leave a constant there with a similar 
behaviour than the electrostatic counterpart l_b, the reaso to do it was in 
case some day one needs to use it for some purpose, and to keep with the same 
"style" of doing things as in electrostatics.  

     I hope the info helps. Best wishes from Mallorca,

         Joan

----- Original Message -----
From: "Markus Gusenbauer" <address@hidden>
To: "Rudolf Weeber" <address@hidden>
Cc: address@hidden
Sent: Friday, September 26, 2014 2:59:49 PM
Subject: Re: [ESPResSo-devel] error in dawaanr interaction

Hi Rudolf,

thanks for your fast answer. When i searched for dprefac*, I couldn't 
find it in the computation of dawaanr. Just with ROTATION, which I don't 
use. Now I am curious again, where does l_b act in the code in case of 
dawaanr? Also when I search for bjerrum, it is also not acting for 
dawaanr interactions.

Markus

On 26.09.2014 12:24, Rudolf Weeber wrote:
> Hi Markus,
> On Fri, Sep 26, 2014 at 10:30:11AM +0300, Markus Gusenbauer wrote:
>> I am wondering if the equation for the dawaanr (dipole-dipole)
>> interaction implemented in ESPResSo is correct
>> (magnetic_non_p3m_methods.c).
>>
>> According to the following papers the force equation in the code is
>> missing the term µ0/(4 pi):
> As far as I'm aware, all dipolar methods don't include this, neither in the 
> force nor in the energy.
> the reason is probably that the pre-factor depends on the unit system, you're 
> using.
>
> You can set the pre-factor, at least in theory, using the Bjerum length
> inter magnetic <l_b> dawaanr
> The prefactor you get in front of forces and energies, according to the code, 
> is (grep -i dprefac.*= src/core/*)
> l_b * k_B T, if temperature >0 and just l_b otherwise.
> So, if you set l_b=mu_0/(4*pi *k_B T), you should get the desired result.
> I'm however not sure, if anyone has ever tested this with dawaanr.
> With the dipolar p3m it was previously done.
>
> Regards, Rudolf
>
>
>
>