Comparisons of energy- and wavelength-dispersive spectrometry inevitably focus on the count rate limitations of the ED electronics. In some cases the use of amplifiers without pulse pileup rejection, or worse, misadjusted pileup rejection circuits, has led to erroneous results and assumptions regarding maximum tolerable count rates which are far too low.

Proper evaluation of pileup rejection must consider both the count rate and energy dependence of the circuits. For example, using a typical ED spectrometer (EDAX, 154 eV, 10 mm2, 6 μsec amplifier time constant) we can first measure the total amplifier throughput curve as shown in Figure 1. This curve is only slightly dependent on photon energy, at least over the range from about 1-10 keV which usually contains the major part of the X rays, and shows a maximum stored count rate of slightly over 3000 cps for an input count rate of slightly less than 10,000 cps. Changing the amplifier shaping time constant will change the scale of the curve, but not its general shape.