Relative move in pin tip center should use setpoint

[DominicOram]

In converting the smargon to ophyd-async we had some issues creating a simulated motor that sets the readback to the setpoint. This made testing the relative move in the pin tip centring plan hard. We should do the relative move from the setpoint rather than the readback, this will simplify the tests and has the same behaviour in prod.

Acceptance Criteria

• The relative move in the pin tip center plan will move based on setpoint rather than readback
• Tests still pass

[DominicOram]

Closing as it's too messy, we should just do bluesky/ophyd-async#57 instead