Currently most components that can be designed scale in their research cost with size, which doesn't make much sense. Highly miniaturized components are difficult in design, while very large components are not necessarily difficult. It is particularly absurd for magazines, why should a large magazine be more difficult to design than a small one, especially when they can simply be stacked?
My suggestion is that the RP cost scaling in size for research projects are changed to either a flat constant or a square root function.
A second point is that currently I can't see any advantage for large magazines/power plants, as many small ones can simply be stacked, thus I suggest that they get some sort of small bonus similar to the increased fuel efficiency of large engines to make them more viable.