It ultimately boils down to how soon you really need to know that the flag is not turning AND how fast the flag is turning to begin with.
One big question is, where is the detector and flag relative to the particular device that is being subjected to the zero-speed question?
grobertson gave no indication as to the relative location in the process.
The fact that one pulses at 90 Hz while the other reads at 50 Hz means that the two effects will "migrate" across each other eventually.
It is certainly true that if the pulse is of a short enough duration the pulse might not be detected for quite a while. However, if the pulse is mechanically extended by making the flag 180-degress (or some such) then the occurance of the cross-over (migration) and detection becomes that much more detectable. Again, the question really is, how soon does the zero-speed really need to be detected?
And again, the answer to the question depends on the location of the pulse generator.
So...?