It really depends on the supervisor, which is a bit of a weird thing since so much depends on a single person - my supervisor let me do anything I want as long

It really depends on the supervisor, which is a bit of a weird thing since so much depends on a single person - my supervisor let me do anything I want as long as there were a few publications at the end (graduation requirement, and in academia the only way to get jobs), other supervisors just look for cheap work for their own project (~robots). I have a friend from the US who worked full-time for 7 years on her PhD and published only one paper! The rest was work on the supervisor's stuff. I would've switched supervisor after 2 years at the latest (but hindsight, 20/20, yaddayadda)