the total mechanical energy of the system is was changing. euler integration
has an energy drift problem for
second order systems.
i've switched to verlet now. two body orbits are mostly stable. total energy fluctuates, especially at high velocities, but is stable over time.