One thing to bear in mind with winglets is that they produce the same effect as additional dihedral, unless they extend equally above, and below, the wing. This can be good, or bad. The chief value of winglets, in my view, is their ability to reduce induced drag within the confines of a given wingspan. Wind tunnel tests revealed that the optimum angle for winglets was pointing straight out - in other words, simply adding to wingspan. Winglets, of course, would produce less additional bending moment in the wing structure than a wingtip extension.

An old NACA report is available on the NASA website, of wing tip plates. These were found to considerably reduce induced drag at high lift coefficients, but total drag was increased up to a lift coefficient of around 0.5 because of the profile drag of the tip plates. I found that tip plates tightened up the loop radius of a competition fun-fly model, and - a bonus - they improved overall handling, since they greatly increase yaw resistance.