 
 
 
 
 
 
 
  
This keyword instructs  GraphEnt to exclude all reflections with 
F/ (F) < f from the 
calculation. The exclusion takes place after all processing of the data is finished 
(ie, taking differences, squaring, etc). It would appear that the presence of this
keyword defeats the purpose of  GraphEnt. This is correct with one exception : When calculating
a 3D isomorphous difference Patterson function, and because the phases of
(F) < f from the 
calculation. The exclusion takes place after all processing of the data is finished 
(ie, taking differences, squaring, etc). It would appear that the presence of this
keyword defeats the purpose of  GraphEnt. This is correct with one exception : When calculating
a 3D isomorphous difference Patterson function, and because the phases of  and
 and 
 are not correlated, a small value for 
(FPH - FP) (for acentric reflections)
will only contain valid information about FH for only half of the cases. Excluding
small differences from this type of calculation is not as harmful as it would be in other cases. My
experience is that giving a  LIMIt 1.0 gives almost identical difference Patterson
functions, but in a fraction of the time that the proper (no exclusions) run would take. In some
cases adding this keyword makes the difference between achieving convergence and 
wasting CPU time.
 are not correlated, a small value for 
(FPH - FP) (for acentric reflections)
will only contain valid information about FH for only half of the cases. Excluding
small differences from this type of calculation is not as harmful as it would be in other cases. My
experience is that giving a  LIMIt 1.0 gives almost identical difference Patterson
functions, but in a fraction of the time that the proper (no exclusions) run would take. In some
cases adding this keyword makes the difference between achieving convergence and 
wasting CPU time.