How do I verify dry powder blend uniformity, and how many sample points do I need?
Do not judge uniformity by eye. Use one sensitive indicator, multiple sample points and RSD.
Pick the indicator that responds most to imperfect mixing: colour difference where the blend is coloured (within-lot ΔE of 1 or less is a common target), whiteness or oil absorption where it is not, and a trace element as a tracer if you want an independent check. Take at least six to ten points — two or three each from the top, middle and bottom, plus the four corners and the dead spots where material accumulates: the discharge, the wall, the area around the shaft. Release at an RSD of 5% or below.
Set the mixing time with an over-mix trial rather than habit. Sample at 3, 5, 8, 10 and 15 minutes, plot RSD, and find where the plateau starts. Mixing past the plateau buys nothing and starts to segregate: light platy powders such as silica and boron nitride float upward, so RSD climbs again the longer the batch runs.
One process rule helps more than any parameter: heavy first, light second. Pre-blend the dense component — barium sulfate, titanium dioxide, iron oxide — with a small portion of the light powder at about 1:3, then add that premix to the main batch. Doing it the other way, all the light powder first, leaves agglomerates against the wall and in dead spots that no amount of mixing will break up.
Finally, a change of machine, speed or fill level means repeating the trial. The plateau time is not transferable between mixers, so old parameters cannot simply be reused.