Research On Nuclear Readout Electronics Using Time Over Dynamic Threshold

Common nuclear radiation detection usually consists of three main parts. Firstly, detector will translate particles information into electrical signal. Then the digitization of the signal will be applied after front-end analog process. Finally, digitalized data will be sent to host computer to be recorded and analyzed.The detector output processing electronics, usually known as the detector readout electronics, plays important role in overall procedure. However, massive channels in parallel are serious burden for readout electronics for most to high-energy physics experiments and nuclear imaging devices. For example, PET detector based on continuous crystal needs to readout 8*8 PMT channels for the light distribution of information a single gamma photon, so that positioning algorithms, such as neural network localization algorithm, SBP algorithm, K-Nearest Neighborhood algorithm and maximum likelihood localization algorithm), can be applied to estimate the event position.For a PET system, which consists of hundreds of detectors, the space reserved for the front-end electronics tend to be very limited while it requires high accuracy performance at the same time. Thus front-end electronics must be highly integrated. Traditionally, charge distribution read method and delay line readout method are used at the loss of resolution accuracy and the readout system formed by standard nuclear signal digitized instrument will require more space, power consumption and money that will exceed the acceptable investment. Up to now, many teams have started studying the multi-channel detector readout methods from different aspects. Based on the objects to be digitalized, methods can be divided into magnitude-based digitalization and time-based digitalization. For amplitude digitization technology, there are mainly three methods:(1) Analog compression, after the amplitude signal is compressed with the analog circuit, it is then digitized by ADC. Though the number of ADC is reduced, it requires specially designed algorithms to support analog compression which still will lose part of the original signal information, thus limits the scope of application of this method. (2) Signal switching, by switching the analog signal from different channels, a high-speed ADC can be used to digitize waveforms from several channels, so that the number of ADC can be saved. However, this method requires higher speed of the ADC. As it is widely acknowledged that the higher speed ADC has, the higher price and power consumption it will cost. (3) CCD based readout, as CCD is broadly used its cost and size can be well controlled. However, CCD must be serially readout at low speed. So this method is suitable only for applications with lower event rate. At the same time, this methods often need wavelength conversion technology, which will give rise to system complexity. For time digitalization technology, there are also mainly three methods:(1) the traditional Wilkinson AD solution:by measuring the time of a nuclear signal over certain threshold, the measured signal is digitalized with time information. Signal is often reshaped by analog circuitry and transformed into a linear discharge form. After compared with a fixed threshold, the nuclear signal is digitized by measuring time through a FPGA-based TDC. However, this technique requires more complex additional analog conditioning circuit, not suitable for massive-channel readout applications. (2) TOT (time over threshold) technology, this method has a simple structure and both the integration and the cost can be well controlled. It directly compares shaped signal with a fixed threshold and measures the over threshold time to recover amplitude. However, the relationship between the time and the amplitude of the signal is non-linear. Though raising the threshold value can reduce non-linear effect, it also reduces the dynamic range of the readout system. Meanwhile, noises can have great impact on the performance of this method, which is usually considered as the most critical problem of TOT. (3) Based on the idea of TOT scheme, we proposed TODT (time over dynamic threshold). Compared with TOT, TODT can be used to digitalize signals with a theoretically strict linear relation between amplitude and digitalized result regardless of the type of shaping circuits, such as RC molding, forming Gaussian, etc.This paper theoretically explain how TODT works with related simulation and test and present how TODT method is applied in nuclear signal digitization. A practical use of TODT is also introduced which offers a solution for the readout system of 64 channels high resolution PET detectors based on monolithic scintillator block and position sensitive photo-multiplier tube to verify the performance of TODT method. As TDC is one of the most important part of TODT, this paper also presents a FPGA-based high-performance TDC, which has satisfying result in logical resource saving and steady performance against changing temperature. Meanwhile, several method to realize dynamic threshold (one digital-based and two analog-based circuits) are mentioned.