The pore-forming α-toxin of Staphylococcus aureus embedded in the planar lipid membrane was used as a nano-device for measuring the size distribution of the degradation products of diamond nanoparticles. As a standard, poly(ethylene glycol) molecules of defined molecular weights were used (according to ref. 1.). We obtained the broad distribution of transient conductance states corresponding to interactions of individual particles with interior of the α-toxin pore. Our goal was to resolve the dimensions of individual particles at low concentration. For recordings with low number of passages (<1000) through one individual pore we faced a trivial problem of histogram construction: we were unable to distinguish properly the individual components. While the sorting of the data into individual histogram categories (bins) led to the loss of resolution when the bin size was to large, decreasing of the bin size led to reduction of histogram amplitudes and to inapplicability of subsequent curve fitting. In this work we present an algorithm which preserves the exact data position in the histogram, leads to improved resolution and allows the deconvolution of histogram components at low number of observations.