Count rate is often determined as the total number of counts collected by the hardware, divided by the live time or acquisition time. One drawback of this approach is the fact that this count rate is not only determined by the environment, but also by noise in the lower channels and by temperature influences. One nice example is shown in the figure below, where a large MS-4000 CsI detector was located in a stationary situation. With the environment not changing, a constant count rate could have been expected, but this appeared not to be the case:
The reason for this change in count rate was a combination of a change in temperature and the use of a lower level in the hardware to cut off counts below a certain level. When the temperature increased, the whole spectrum started to drift to the right, and more gamma particles are registered above the lower level, thus increasing the registered count rate.
When applying the Medusa energy stabilization algorithm, the drift in the spectrum is compensated for, and the count rate based on the stabilized spectrum is much more stable:
Bottomline: Always stabilize your spectrum before calculating the count rate.