From my research, the mechanical fuel pumps do not produce pulsing on the pressure side, BUT DO produce pulsing on the suction side. Input check valve is open on the pump upstroke and closed on the down stroke, creating the starting and stopping of flow with the slightest amount of backflow as the check valve closes. I believe this why the flow sensors are recommended to be installed downstream of any pumps. Haven't looked at floscan, but EI definitely makes this clear for the red cube in their documentation. The constant starting and stopping of flow on the suction side makes the sensor go wonky. It is counting the vanes moving by and the starting / stopping makes this very difficult to smooth out, as the vane can reverse a bit when flow stops and can be counted twice if the vane stopped right at the sensor and reversed a bit. The sensor is not directional, it just counts vanes moving past and therefore, reversals are not accounted for. The red cube is optical, but don't know if the floscan is optical or something else. Optical is VERY binary and the slightest fraction of movement in reverse can result in double counting a vane IF the van stopped very close to the sensor.
I suspect in your case, the fluctuations stop when the boost pump is on (iow, sensor now on the pressure side).
Both of my planes have the cube downstream and flow is solid as a rock in cruise.
Larry