In practice the 4046 PLL for a Nipkow disc with a missing sync pulse will synchronise one line shifted. I see this in the MUTR televisor and I have seen it in the Peter Smith and Grant Dixon monitors as well. If the phase difference is more than one line the correction pulse is so long that the disc will slip one line in a revolution. But with one line offset the corection pulse is to short.
With correction by hand (gentle braking) it is possible to synchronise properly. The correction wave form is perfect in this case and the disc keeps its synchronisation. But when you let it go from not synchronised situation it will most of the time come to one line offset. This can be seen best on the Grant Dixon test pattern, where left and right should show equal 2 lines width grey scales. If one line is swapped from left to right or reverse you will see that clearly.