-when you get down to the 70u pile it all into a pile and get a slip of white paper nice and clean. scoop up a small small amount that would fit under the microscope. look at this under the microscope (or loupe if it works) and if you use a usb one take a picture or memorize and note the head:contaminate ratio. put the kief back on the 70u to clean more.

-use the same method(technique) your trying to evaluate and clean on the 70u then after 5-10 minutes of gentle carding using the SAME method to eliminate variables. scoop a small amount up again with clean white paper again check under scope the head:contaminate ratio is it better or worse? it should most likely be more heads and less contaminate.
-repeat this till clean(if ratio improves each time magnified) OR change technique till it cleans it faster etc.

i also should add that when im cleaning the kief on the 70u ive noticed lately without magnifying that its usually done when the pile stops shrinking or starts shrinking very slowly. it has taken me many cleanings to be aware of the rate at which kief contaminate ie capitate stalks etc fall through(i still magnify). but the pile basically stops shrinking(shrinks real slowly) when the heads are left. hopefully soon ill make a breakthrough and be speed cleaning the sift scalable to large amounts haha once i do ill let you all know how if i get it super fast peace