increasing the accuracy of LOCPOT

[user224]

Hi all,

I've been trying to calculate accurate absoulte band energies in order to compare the effects of straining a slab structure. To do this I've been using the usual method of LOCPOT plus the vtovav script. While the local potential does converge, there seems to be a reasonable amount of 'noise', with oscillations of up to 1 meV once convergence is reached. I've tried to reduce this noise by increasing ENCUT and also explicitly increasing NGX,Y,Z, but with little success.

1 0.31213011742E+01
2 0.31220002174E+01
3 0.31217055321E+01
4 0.31218757629E+01
5 0.31218075752E+01
6 0.31218242645E+01
7 0.31220278740E+01
8 0.31220154762E+01
9 0.31222577095E+01

Does anybody have any tips regarding increasing the accuracy of LOCPOT in general?

Any help would be very much appreciated.

Thanks

[nwan]

Have you checked maximum and minimum number of planewave?