Jonas, Two things. First, if you have a fixed call calculation, there is no need to specify cellside and cellangle, since they are not implemented yet (as stated in the manual!). Just set MD.VariableCell to false and all is set for a fixed cell.
Looking at your input, I realize you separate the atom positions by a tabulation character. A wild guess: maybe this is giving the parser a hard time and it is not setting all atoms fixed? What if you remove the tab and simply use a space as a separator to the atom indices? You are correct, your positions should not change if you set position to fix them - unless you state that they are constrained only in specific directions. On a superficial look, your syntax seems to be correct, so this is the only reason I can think of at the moment. Marcos On Tue, Apr 27, 2010 at 3:59 PM, Jonas Baltrusaitis <[email protected]>wrote: > Dear all, > > after trying several times I have to conclude that geometry constraints in > Siesta do not work. In the attached output (and several others) I tried to > constraint bulk SiO2 in atomic position while relaxing only OH groups. To my > surprise, all of the atoms moved, though inner, constrained atoms, somewhat > less. Still, I was thinking that constrained atom position should not change > at all. Does constraint work differently in Siesta and it allows those atom > to relax? > > Jonas > >
