There seems to be quite a lot of speculation that the autofocus performance on the Z5 must be different than that of the Z6. Without a camera in hand, there's no real way to debunk that theory, but let me explain a couple of things and where the impacts would be, then state my hypothesis.
In Nikon's dual pixel design used on every Z system camera so far, every twelfth row of photosites on the sensor has a phase detect "mask" associated with it. That mask lives in the layers above the actual sensor silicone (e.g. microlens layer). The Z5 appears to be no different in that respect.
What might be different is how fast data in those rows are moved off the sensor. While we don't know the origin or fabricator of the sensor, somehow I doubt that this would be true, or would be significantly different, on the Z5. Why? Because it appears that to save costs, Nikon is reusing the same downstream components, including the components on the digital board that would receive that data and act on it. EXPEED6 is also the same. Moreover, the EVF on the Z5 is the same as on the Z6, indicating that same speed of sensor data offload is happening.
So while there might be some slight timing issues that are different within the sensor, I'll bet that EXPEED and its brethren chips are getting essentially the same amount of data in the same time frame.
My hypothesis? Z5 autofocus will look amazingly like Z6 autofocus.
