Very few concern themselves about the strength of solder connections, because it doesn�t require much strength to keep a solder connection attached to the board. More important are factors that arise as a result of the different materials properties of the parts and boards that go into an assembly, such as: * Stress / Strain relationships. * Creep characteristics.
But we digress.
�Since all of them� [the 4-5 different land dimensions???] �were� [life cycle???] �tested� [in end use environment???], the board will never know you have different pad designs on it. Don�t go back and standardize them. Let sleeping dogs lie.
Using multiple pad designs for the same component seems to be a clever conspiracy by board lay-out types to drive production types nuts. The lay-out types explain this with some convoluted mumbo jumbo about careful design considerations given the board performance characteristics, prepreg and laminate selection, supplier recommended patterns, the early onset of senility, and �I sure wouldn�t have done it this way. The designers don�t know what they�re doing and they made me do it.� [Actually, even with those gobblie gook laden explanations, we production types think the lay-out types just don�t know what they�re doing. In their defense, we grudgingly will admit that often by the time lay-out types get a job, the designers have done obscene things to the schedule and the pooch.]
In helping to explain the reason for the myriad of different land patterns, different end-use environments require different amounts of solder on the connection and so, use different pattern designs for the same component. You can experiment with this using the IPC [www.ipc.org] land pattern calculator. It is based on the SM-782 - Surface Mount Design & Land Pattern Standard.