The flow fluctuation characterization device (FFCD) consists of two small vessels each open at the top and having an orifice at its lower end. The vessels are mounted one above the other, with the lower vessel attached to a load cell or other means for continuously weighing its contents. The material under scrutiny is admitted to the upper vessel, flows into the lower vessel, and exits via the orifice in its lower end. The orifice in the lower vessel is smaller than that in the upper vessel so the flow is choked at this point and the lower, weighed, vessel is kept full of the test material. It has been found experimentally that the load cell output fluctuates with time, and is sensitive to the physical properties of the test material such as particle size and size distribution, the presence of small quantities of fines, and the relative proportions of the species in a mixture. This paper considers cases where the differences in the physical properties in batches of test material are small, and discusses the application of signal-analysis techniques to the time series produced by the FFCD, to establish with confidence that the samples can be considered to be different. Signal-analysis techniques including spectral analysis and attractor reconstruction are critically evaluated.