A common situation is the case when
the signal consists of a series of pulses of essentially identical shape corresponding to a series of events. The
pulses are related to each other by a random scaling factor so that significant information is contained in the
height of the pulse, h, and how this quantity is distributed. The latter is described by the probability function
p(h) defined so that the probability of occurrence of a pulse with height between h and h+dh is
In cases when the pulses are of microsecond
or less duration the time for which the signal is approximately at h is extremely short so that special steps must
be taken both to hold the height for analysis and to detect that the maximum has been reached. An arrangement to
capture the peak height is based upon the sample and hold procedure. In principle the task could be accomplished
by issuing the hold command the instant the maximum is reached. This approach is unsatisfactory since the timing
requirements are too stringent. A simple addition to the circuit as shown in the diagram provides the solution.
The main feature is the inclusion of the diode at the sample and hold input. During the rising portion of a signal
at this input the current flow is forward and the diode conducts charging the capacitor, and the voltage across
the latter follows the input. At the moment the peak voltage is passed the diode is back biased, so the peak voltage
is held on the capacitor. The ancillary circuitry indicates some of the necessary interfacing to the ADC.
The control FF is set by the ADC end-of-conversion signal, EOC, opening the linear gate. The peak detector PD resets the control FF closing the gate, and signalling the ADC to start a measurement cycle. The gate remains closed until processing of the currently sampled pulse height is completed at which time an EOC is issued. The PD performs a role analogous to the internal trigger of an oscilloscope, indicating the occurrence of a signal at the input, as well as determining the time at which the peak is reached.
The principle of peak detection is
illustrated in the accompanying figure. The input signal is differentiated. The derivative signal now crosses zero
at the time the signal becomes maximum.
The cross-over point is detected using a Schmitt trigger. This is designed so that the upper level and hysteresis value are identical so that the lower trigger level is precisely zero. The trigger fires when the input exceeds the upper trigger level H, returning when the input reaches zero. Peak time is then determined from the negative going edge of the Schmitt trigger output as indicated.
Time mark generation by cross-over detection is a generally important technique since the cross-over point is independent of the pulse height.
Once the conversion is completed the ADC contents are transferred to a memory address register, and the memory contents of the location at that address are incremented by one, indicating the occurrence of an event with height in the interval associated with the address. If a measurement is continued for a time T, during which period a total of NT events are processed, then the expected content of memory location n is
The data constitute a histogram approximating the probability density function if the quantizing voltage is sufficiently small so that the integrand may be approximated by a constant value p(nv0) over the integration interval v0.
This type of digital measurement system is referred to as a multichannel pulse-height analyser and has many applications. Three examples will be discussed. The first is connected to the technique of flow cytometry in which individual cells in suspension flow through a capillary. A fine beam of ultraviolet radiation is directed across the diameter of the capillary so that a cell passes through it as it flows along. The ultraviolet light causes a specific molecular species such as DNA to fluoresce at a characteristic wavelength and the fluorescence is detected by a photo-detector with the appropriate spectral response. The intensity of the florescence is determined by the DNA content of the cell and in turn determines the magnitude of the pulse from the photo-detector induced by the cell passage through the UV beam. The photo-detector output is coupled to a multichannel analyser so that the distribution of DNA content in the cell population may be determined.
The second example employs a device known as a Coulter counter to determine particle size. The counter consists of cell (as in dry cell - not biological cells of the prior example) containing an electrolyte. A pair of closely spaced electrodes with the electrolyte in the gap forms a resistor. Particles are suspended in the electrolyte and a flow is maintained. Diaphragms direct the particles individually to flow through the electrode gap. As the particle traverses the gap it displaces the electrolyte and increases the cell resistance. The amount of electrolyte displaced and hence the cell resistance is a function of the particle size. With the cell forming part of a divider chain connected to a voltage source the resistance change is converted into a pulse the height of which provides information regarding particle size. An important application is the size distribution of pigment in paints.
The third example is the application for which the multichannel analyser was originally developed, nuclear pulse spectrometry. An energy sensitive radiation detector responds to an interaction with a quantum or particle of radiation by issuing a pulse the height of which is proportional to the energy deposited in the sensitive material of the detector. The pulse height distribution acquired in a multichannel analyser, usually referred to as a pulse height spectrum, provides information regarding the energy spectrum of the radiation.