How much weight are we talking about here, grams, kilograms? You could use a laboratory-scale end-over-end drum tumbler (with e.g. rubber stoppers put in the drum) which may help to thoroughly mix the two things.
You have a classic mixing or sampling issue dependent (this is just one one factor) on the particle size distributions differences in the samples. You don't tell us about the particle size distributions in the samples. The large particles will govern the minimum mass required and you need to decide your level of homogeneity (really heterogeneity). Do you want your finished mix homogeneous to a 95 or 99% confidence level? From a practical perspective the means of the 2 materials should be as close as 0.7 - 1.3 to minimize segregation. For calculation of minimum masses see (for example slide 26 and following in the first url):
December 14th, 2010 Sampling for particle size analysis - estimation of standard error