[gmx-developers] dispersion correction / exploding box

[Rainer Böckmann]

...the simulations differ only in the number of water molecules / the 
box size. The protein topology works fine... A smaller box with d=1.2nm 
using the same topology yields a negative dispersion energy and constant 
volume.
Best,
rainer

David van der Spoel wrote:

>On Tue, 2005-08-23 at 09:35 +0200, Rainer Böckmann wrote:
>  
>
>>ok - c6 is negative...
>>
>>Average C6 parameter is: -0.00162519
>>Average C12 parameter is: -1.70729e-06
>>
>>    
>>
>
>then your topology is weird. 
>
>  
>
>>Berk Hess wrote:
>>
>>    
>>
>>>Rainer Böckmann wrote:
>>>
>>>      
>>>
>>>>Hi Berk,
>>>>
>>>>yes, for certain box/system sizes I get a positive dispersion energy 
>>>>resulting in drastically increasing box sizes.
>>>>
>>>>        
>>>>
>>>But that seems impossible, unless you have positive c6 parameters.
>>>Could you run with the -debug option and check what the average c6 
>>>parameter is?
>>>
>>>Berk.
>>>
>>>_______________________________________________
>>>      
>>>
>>    
>>