A new model of clinical judgment
The purpose of this investigation was to study the application of a refined constant-fraction discriminator to amplitude-and-risetime compensated (ARC) timing with a Ge(Li) detector. Three distinguishing features of this ARC timing discriminator were discussed. First, a lower-level leading-edge (LLLE) discriminator is used to arm the constant-fraction zero-crossing detector. The threshold level for the LLLE discriminator is automatically set at one-half the minimum energy level required to produce an output pulse from the unit. An upper-level leading-edge (ULLE) discriminator determines the minimum energy level of interest and arms a coincidence gate for the detection of the delayed logic pulse from the zero-crossing detector. The second distinguishing feature of the instrument is the delay tk, in the path of the timing logic pulse from the crossover detection device. This delay allows approximately 19 nsec. after zero-crossing for the input signal to trigger the ULLE discriminator. Input signals not meeting this constraint produce leading-edge timing at the coincidence gate. The third distinguishing feature of the ARC timing discriminator is the addition of a Slow-Risetime-Reject (SRT) circuit, which can be employed to reject the logic pulses resulting from leading-edge timing. This SRT circuit is intended for use with Ge(Li) detectors or other similar semiconductor detectors, which have charge collection times ranging from 20 nsec. to 200 nsec. or more.
Criteria were derived for the constant-fraction shaping delay, td, and for the fixed circuit delay, tk, to insure ARC timing and to permit the SRT circuitry, when employed, to eliminate logic pulses resulting from leading-edge timing. These criteria were established for three specific input signal shapes-a decreasing slope signal, an increasing slope signal, and a constant slope signal.
Timing resolution improvements resulting from the use of the SRT circuitry were measured. A gamma-gamma timing coincidence experiment was performed, using a 1" x 1", cylindrical, KL236 plastic scintillator, on an RCA 8575 photomultiplier tube, and a true coaxial Ge(Li) detector. The Ge(Li) detector is a right circular cylinder, with a 58.6 cc active volume and a 10.3 percent measured relative efficiency. Timing resolution data were obtained for coincident events in a 1.1:1 dynamic range of energies around the 0.511 MeV energy of 22Na. Timing resolution as a function of threshold level was determined both with and without the use of the SRT circuitry and was plotted. As an example of the resolution improvement, use of the SRT circuitry provided an improvement in the FWHM timing resolution of 33 percent for a 350 keV threshold setting. An improvement of 48 percent was realized in the FW(1/10)M resolution for a 250 keV threshold setting. Photographs of timing spectra qualitatively showed the dramatic improvement in timing that was obtained with the SRT circuitry.
Timing resolution in the Ge(Li) timing system was improved by using the SRT circuitry; however, the counting efficiency was reduced. Because of the risetime constraints imposed by the SRT circuitry, some input signals that have sufficient amplitude to exceed the threshold level of the discriminator do not produce output pulses. Analytical expressions were derived to describe the relative loss in the counting efficiency of the discriminator when the SRT circuitry is employed.
The expression for the loss in counting efficiency in a particular energy peak as a function of the relative threshold level was verified experimentally. Energy spectra were gated by the ARC timing discriminator, and the counting efficiency loss was determined for the 0.511 MeV photopeak of 22Na and for the 1.174 MeV photopeak of 60Co. The calculated and measured values of the threshold level at which counting efficiency loss became total differed by less than 7 percent. Photographs of several gated energy spectra qualitatively illustrated the loss in counting efficiency caused by using the SRT circuitry.
Calculations for the loss in counting efficiency of the discriminator required a knowledge of the risetime probability density function of signals from the Ge(Li) detector and amplifier system. An experimental technique for measuring this distribution of risetimes was described. For an uncollimated gamma-ray source, the risetime probability density function of signals from the coaxial Ge(Li) detector and amplifier system was shown to be a decaying exponential with a 32.6 nsec. average decay time constant. The distribution was approximately independent of energy in the range considered.
Several additional analytical expressions were derived to further characterize the effects of pulse shapes on constant-fraction timing. An expression was developed to relate the noise on the constant-fraction signal to the noise on the input signal. Expressions for noise-induced time jitter and for zero-crossing times were derived for true constant-fraction (TCF) and for ARC timing with the three previously mentioned input signals.
Thesis77b.B454.pdf
2.61 MB
Adobe PDF
ec984a6fb84cf3b1c9c173d92739d305