The display update was 1hz. I am dealing with mains pressure water with straight pipe 200mm each side. The actual flow would be about 5 l/min but the readings are all over the place 2.13, 6.29, 4.96, 3.45, that sort of stuff, and by the time you have understood one reading that is so different from the one before, you have missed the next two. Not that that matters, as you have no idea which one to use anyway.
The problem was not so apparent with my Swissflow axial flow turbine, but that was using the frequency meter on a DMM, I never used it with Arduino. I did have a bad time with it, the problem may actually have been the same, and it may be a characteristic of all hall-effect turbines.
I have not progressed any further with this but rate is vital for me and I have to resolve it. I'm sure it can be fixed by either averaging ten readings or simply making the count period ten times as long, or something like that.