Abstract
The set of large Es through which a structure is solved by direct methods is usually chosen by a convergence or convergence–divergence process. This process aims to give a strong phase-extension pathway starting from a small set of Es whose phases are known or allocated in some way. Sometimes sets of reflexions thus obtained are poorly conditioned and under tangent-formula refinement even initially correct phases will degenerate to randomness. A simple new algorithm has been developed which improves the conditioning of the complete set of reflexions and their relationships and is more appropriate to current trends to start refinement from a complete set of random phases. A particular feature of this algorithm is that it maximizes the minimum number of relationships for any reflexion.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
