Volumetric whole bean coffee dosing, a quick consistency test

#1: Post by **Bluecold** » September 14th, 2015, 12:51 pm

As the Versalab uses a volumetric bean-dosing system, and I had a doser laying around gathering dust, I figured I could repurpose my doser to dispense precise amounts of whole beans. This would save me some mucking about with a scale as my grinder is basically retention free.
So I threw some beans in my doser, and measured the consistency of the output when thwacking the doser.
I used one-thwack and double-thwacking and set the doser at two heights, very low and as high as it could go. I did twenty doses with each experiment run. Each time I dropped the dosed beans back in the doser for a consistent doser height. The doser was half-full and completely covered the dosing star. The first three samples where not counted with the experiment to ensure consistent filling.
The results are as follows:

-----------------------------------------------------------------
						 low height			|		higher height	
-----------------------------------------------------------------
				1 thwack	2 thwack		 |	1 thwack	2 thwack
-----------------------------------------------------------------
Mean (gr)	5.01			9.78			|	9.04		18.56
Max  (gr)	5.3			10.7 			|	9.4 		19.5
Min  (gr)	4.5			 9.3 			|	8.7 		17.8
StDev(gr)	0.22			0.35			|	0.25		 0.50
Variance    0.05			0.13			|	0.06		 0.25
-----------------------------------------------------------------

The results are a bit puzzling as one would expect the double-thwack experiments to have less standard deviation and variance as you use two samples. Apparently, there is a bit of correlation between the samples. And I expected a higher bean column to dose more consistently as the relative impact of a few beans being scraped away from the top would be lower. In any case, I think this is a solid indication that volumetric bean dosing is not sufficient for the discerning home user. Or that a doser designed for ground coffee is not suitable for whole beans. What the parameters for consistent volumetric whole-bean dosing would be I don't know.

--edited to clarify the reason for the test---
For those wondering about the consistency of a doser for ground coffee, Jim Schulman stated this to be within 1/3 gram of the desired dose
How I Avoid Single Dosing