EN
Fast and slow channels
The fast and slow channels are two complementary signal processing capabilities within the digital pulse processor of Amptek detectors. They operate at different peak times, which optimizes the information obtained from the examination of different sets of pulses.
The slow channel, which has a long peak time, is optimized to eliminate electronic noise and "ballistic deficit" for accurate peak height recording, free from extraneous noise.
The fast channel, which has a short peak-to-peak time, is optimized for detecting pulses recorded at approximately the same time (closely spaced in time) that overlap (accumulate, increase) in the slow channel.
For most detectors, electrical noise is eliminated at fairly long peak times. Amptek's Si-PIN XR100 detector achieves minimum electronic noise with a peak time of about 25 µs, for SDD the minimum is reached at 10 µs. Using this detector, the peak time for the slow channel can be adjusted in the range of 1 - 20 µs. Its output is connected to the peak detection circuit and is used to measure the energy spectrum.
Since the radiation in the detector travels different distances, it is possible for two interactions to occur in less than the slow channel time, so that the two events are registered as one. This is possible even at low count rates. In most studies, it is useful to use high count rates, so overlap will occur frequently. There are two problems with pulse overlap in time:
- Instead of two clearly resolved signals, one is recorded.
- The peak magnitude is distorted and does not carry information about the real interaction.
The graph below shows the following results:
(A) This is a single non-overlapping signal that is registered correctly with the correct amplitude.
(B) Two pulses that have a slight overlap, but their amplitudes are recorded correctly.
(C) Two pulses, the interval between registrations for which is less than the peak reaching time in the slow channel, will register a single event with incorrect amplitude. For this case, the dead time Tdead ≥ Tpeak + Tflat . Where Tpeak is the peak reaching time in the slow channel.
To solve this problem, DPP has a fast channel with much shorter peak-to-peak time. In the latest DPPs, this time can be set in software (maybe shorter than 50 ns). The red trace in the figure is the fast channel trace. Many pulses that overlap in the slow channel can be resolved in the fast channel. The fast channel is used to measure the exact count rate of events, in which two pulses are resolved before they overlap in time. This is used in the pile-up rejection (PUR) logic. When this logic is enabled, the event (C) will be rejected to avoid registering pulses with incorrect amplitudes, which will only distort the obtained spectrum. No events are registered in the slow channel, although they are displayed in the fast channel. The figure above shows another important and more delicate point: the fast channel has more noise than the slow one. Different thresholds are used in the fast and slow channels, so the common threshold can be set higher than the common noise.
The figure below shows another important difference between the fast and slow channels. Simulated response for three different pulses. These pulses have the same energy, but the rise time of these pulses by the preamplifier is 0.2 µs, 0.8 µs, 1.6 µs for (A), (B), (C) respectively. In the slow channel Tflat is 0.8 µs, so the amplitudes of peaks A and B are equal, and the amplitude of peak C is smaller by a barely noticeable amount. The dependence of the peak amplitude on the pulse rise time is called the "ballistic deficit", and together with the trapezoid shaper used in the Amptek detector, the ballistic deficit appears if and only if the peak rise time is greater than Tflat. In the fast channel, the peak amplitude mainly depends only on the peak rise time. This is the basis of the discrimination ramp in DPP. The ratio of the slow and fast channel pulses is related to the rise time.
