Questions about crosstalk or spillover are frequently asked by users of alpha/beta counting systems. There is some confusion regarding the definition. Technically (in an engineering sense) crosstalk means a transfer of signal from one channel of a system into another channel of that same system. It is sometimes used in a multi-detector system to refer to a signal seen by a detector other than the target detector. Common usage of crosstalk is that it is the overlap of the alpha and beta counting regions of a detector or counting system. The last definition is more correctly referred to as spillover. Manufacturers of alpha/beta counting systems specify spillover as X.Y% alpha into beta, and 0.Z% beta into alpha, for ²¹⁰Po. The last definition and specification is the one discussed here.

Protean Instrument's specification is that all systems with a gas flow proportional detector have 0.0% spill-over, either alpha into beta or beta into alpha. The actual test criteria is that spill-over is no greater than 1 count per 100,000 counts in either direction. Protean is the only manufacturer that can promise and deliver this extraordinary level of performance.

		The problem: older systems, and current systems from other manufacturers have overlapping alpha and beta regions. Pulses falling into the overlapping regions cannot be differentiated as one or the other. The instrument manager must make a decision as to how pulses in the overlapping region are handled.
		The oldest solution as used in classic systems: use a single discriminator to divide the overlap region into alpha and beta. The instrument manager makes an arbitrary decision that determines the threshold for separating alpha pulses from beta pulses. Some pulses will always be counted incorrectly. Some efficiency will be lost in one or both channels.
		The next solution: use two discriminators or an MCA and two regions of interest (ROIs). Both systems allow an arbitrary dead band to be set between the beta region and alpha region. Both methods discard valid beta and valid alpha counts.
		Protean's "loss-free" method: use a detector with a wide dynamic range. The detector completely differentiates alpha and beta. No alpha counts are in the beta region and no beta counts are in the alpha region. No counts are lost.