There is much noise, chaotic behaviours in impact plate method. It is no problem in high flowrates as electronic filters can work well to surprass the noises so to neglect small vibrations, noise, whatever you call, of granular particles. So, error margin in measuring low volume flowrates will be high or, actually, things neglected when averaging at electronic filter will start to become visible more. So, error by impact plate technique goes high as flowrate goes low.

Impact plate method of measuring granular flows is based on free falling of particles, in other words, based on measuring forces created by linear inertial acceleration (remember newton's law, f = m.a where a is vectorial summation of linear, radial and coriolis accelerations.) Impact plate uses linear component, in vertical falling, it is gravitational acceleration, g.)

Theoretically, radial acceleration method will be a little better. There are devices now which uses this principle, usually called chute plate type. There is also coriolis type flowrate measurers as you know in other fields, but, lately, there are some measurement devices for granulars too based on this coriolis method. The best method for lower flowrates can be coriolis method as you can control parameters (rpm and rotor diameters) finely and it'll surprass the noise better at high RPMs. If this doesn't work too, for your allowable error margin, then, you'll think probe-less measurment methods. ■