Can you derive Newton's law of gravitation from Kepler's third law, assuming an elliptical orbit? Most of what I've seen have been people solving it with a circular orbit. However, I find it impossible for an elliptical orbit because the radius changes as one object orbits a bigger one.
I first tried work backward: deriving Kepler's law from Newton's. I ended with the fact there are two components to acceleration (the one working with or against the velocity and the centripetal one), which can be shown as this $$ \vec{a} = v'\hat{T}+v^2κ\hat{N} $$ v and v' stands for velocity and the derivative of velocity, T_hat is the unit tangent vector, κ is the curvature, and N_hat is the normal vector.
I ended up with this $$ -\frac{GMm}{r^2}\hat{r} = m\vec{a} $$ $$ -\frac{GM}{r^2}\hat{r} = \vec{a} $$ $$ -\frac{GM}{r^2}\hat{r}=(v'\hat{T}+v^2κ\hat{N}) $$ Because the velocity does simply equal 2πr/T, like most examples use: $$ \frac{GMm}{r^2} = m\frac{v^2}{r} $$ I, again, don't see how one can derive Newton's law of gravitation from Kepler's 3rd law. I feel like I am missing something; are my understanding and reasoning wrong? or is there a way to derive it?