The influence of charging tube materials and diameter on the separation efficiencies of a gluten‐starch model mixture and lupine flour was studied. Offline analysis of tribo‐charging with different tube materials showed that gluten takes a positive and starch a negative charge. However, the charge of the mixture was found not equal to the sum of the charge of the individual components and measured charges could not be related to the triboelectric series. During electrostatic separation significant protein enrichment was observed for both plant raw‐materials. For the model mixture differences in protein, enrichment were observed between tube materials, but this was not the case for lupine flour. The lupine protein content increased from 37 to 65 g/100 g dry flour. Concluding, electrostatic separation needs to be evaluated during separation experiments, as particle‐particle interactions dominate the charging process and thus separation of mixtures.Practical applicationsThe combination of dry milling and electrostatic separation is investigated as a sustainable and mild route for protein fractionation. The results of this study showed that offline charging tube experiments could not predict separation performance of for example finely milled lupin flours. Instead, performance should be directly assessed during separation experiments, which is explained as charging is rather related to differences in material or triboelectric charging properties between powder particles than to charging tube wall properties. The results of this study benefit development of new applications for electrostatic separation.