阅读说明：本技术 PET的Wave-union型TDC装置及测量方法 (Wave-unity type TDC device of PET (polyethylene terephthalate) and measuring method ) 是由 马聪 李兴 黄振强 于 2020-08-11 设计创作，主要内容包括：本发明涉及时间测量领域,尤其涉及一种PET的Wave-union型TDC装置及测量方法,利用在进位链中引入Wave Union Launcher激发输入脉冲信号产生周期性震荡,通过多次测量信号的前沿时刻达到提高测量精度的目的；另外,还利用晶体本底噪声进行TDC的在线非线性修正,进一步提高时间分辨率和统一数据格式,方便后端符合逻辑算法；优势在于：该装置具有精度高、死时间小、成本低和易集成等优势,满足绝大部分PET系统应用需求。(The invention relates to the field of time measurement, in particular to a Wave-unity type TDC device and a measuring method of PET (positron emission tomography). The purpose of improving the measuring precision is achieved by measuring the leading edge time of a signal for many times by introducing a Wave Union Launcher in a carry chain to excite an input pulse signal to generate periodic oscillation; in addition, crystal background noise is used for performing on-line nonlinear correction on the TDC, so that time resolution and a unified data format are further improved, and the back end can conveniently conform to a logic algorithm; the advantages are that: the device has the advantages of high precision, small dead time, low cost, easy integration and the like, and meets the application requirements of most PET systems.)

1. A PET Wave-unity type TDC device comprises an input end D trigger, a multiplexer, a NOT gate, a carry chain, a Wave-unity Launcher, a DFF array, a pipeline encoder, nonlinear correction logic, fine time calculation logic, a coarse counter and a data buffer;

the input end D trigger is used for latching and clearing an input pulse signal;

the CARRY chain device is formed by connecting CARRY4 hard cores in the FPGA in series, input pulse signals are transmitted in the chain in an incremental manner after entering, and the pulse arrival time is marked by using different tap level states on the chain;

the DFF array is composed of D triggers and used for latching the level on a carry chain tap and generating a thermometer code;

the Wave-unity Launcher utilizes one tap of a carry chain and feeds back the tap to the input end of a multiplexer through an NOT gate to generate periodic oscillation pulse;

the Pipeline encoder is used for converting thermometer codes into binary coded fine time information;

the fine time calculation logic is used for normalizing fine time information obtained by multiple times of oscillation and summing or averaging to obtain a final fine time result;

the non-linear correction logic is to correct a final fine time result;

the coarse counter is an FPGA internal counter soft core driven by a sampling clock and is used for obtaining coarse time information;

the data buffer is used for buffering the obtained time data and communicating with other logics of the system.

2. The PET Wave-anion type TDC device according to claim 1, wherein: the carry chain delay Δ chain should be greater than the sampling clock period Tclk.

3. The PET Wave-anion type TDC device according to claim 1, wherein: the CARRY4 includes four taps, where the level states of two taps, whose delays to the DFF array are consistent and minimal, are used to mark the pulse arrival time.

4. The PET Wave-anion type TDC device according to claim 1, wherein: the Pipeline encoder consists of a one-hot code encoder and a binary encoder; the one-hot code encoder obtains one-hot codes by detecting thermometer codes generated by DFF array latch; the binary encoder comprises a basic tree structure and an output tree structure, wherein the basic tree structure utilizes an OR gate to build a layer tree structure, the output tree structure extracts the middle level of the basic tree structure, and finally the one-hot code is converted into binary encoded fine time information.

5. The PET Wave-unity type TDC device as claimed in claim 4, wherein: d flip-flop arrays are added behind each layer of gates of the basic tree structure and the output tree structure.

6. A method for measuring a Wave-unity type PET TDC device is characterized by comprising the following steps:

s01, inputting a pulse signal to be detected, enabling the pulse signal to be detected to enter a carry chain after passing through an input end D trigger and a multi-path selector, selecting 1 by the multi-path selector when the pulse signal is transmitted to a kth tap, and pulling down an input signal of the carry chain; when the low level reaches the kth tap again, the multiplexer selects 0, the carry chain input signal is pulled high, so that periodic oscillation pulses are generated, and the oscillation period Tosc of the Wave-unity Launcher is greater than the carry chain delay delta chain; after an input pulse signal enters, the input pulse signal is transmitted in a chain in an incremental mode, the pulse arrival time is marked by using different tap level states on the chain, and the carry chain delay delta chain is larger than the sampling clock period Tclk;

s02, the DFF array latches the level on the carry chain tap and generates thermometer code;

s03, the one-hot code encoder obtains the one-hot code by detecting the thermometer code, and the binary encoder converts the one-hot code into binary encoded fine time information;

s04, the fine time calculation logic is to normalize the fine time information obtained by multiple times of oscillation, and sum and average to obtain the final fine time result, the normalization is performed by a recursive operation formula

s05, the nonlinear correction logic is used for correcting the final fine time result;

s06, the coarse counter uses the FPGA internal counter soft core driven by the sampling clock to obtain the coarse time information;

and S07, combining the coarse time information and the corrected fine time result to complete time measurement in a dynamic range.

7. The method for measuring a Wave-unity type TDC device for PET as claimed in claim 6, wherein: in step S04, the fine time information for each measurement is averaged, where the average calculation formula is:

where M is the total number of measurements, t (M,1) is the fine time result relative to the first measurement, t (M) is the fine time result of this measurement, and n (M) is the measurement period function that needs compensation for both beta and gamma type events.

8. The method as claimed in claim 6, wherein the oscillation period Tosc is determined by the following equation:

wherein, Δ tap is the carry chain tap average unit delay, Δ MUX is the multiplexer delay, Δ INV is the not gate delay, and Δ ROUT is the routing delay; the determination of the fractional delay comprises the steps of: and fixing the positions of the NOT gate and the multiplexer in the FPGA by using a development tool, carrying out delay time sequence constraint and routing fixation on the delta ROUT, and finally estimating the delay of each part according to an imitation result after tool wiring.

9. The method for measuring a Wave-unity type TDC device for PET as claimed in claim 6, wherein the step S03 further comprises the steps of: carrying out NAND operation on each bit of the thermometer code and adjacent high four bits of the thermometer code, only marking the position where the highest jump of '01' occurs, and respectively carrying out AND operation on the highest three bits and the lowest bit, wherein the Boolean operation formula is as follows:

。

10. the method for measuring a Wave-unity type TDC device for PET as claimed in claim 6, wherein the step S05 includes: the nonlinear correction logic utilizes random pulse signals generated by background noise of a detector, counts the occurrence frequency of each fine time when the system is started, calculates the delay uniformity, obtains nonlinear correction parameters by utilizing a bin-by-bin method, stores the nonlinear correction parameters in a Random Access Memory (RAM) in a Field Programmable Gate Array (FPGA) as a lookup table, and reads the correction values in the lookup table according to the final measured fine time result as an address to obtain the corrected fine time value.

Technical Field

The invention relates to the field of time measurement, in particular to a Wave-unity type TDC device and a measuring method for PET.

Background

Positron Emission Tomography (PET) based on the Time-of-flight (TOF) imaging principle is favored because of its high signal-to-noise ratio and low random coincidence. The detection principle is to convert gamma rays emitted from a radioactive source into visible light by using a crystal (such as Lyso), and then convert the visible light into an analog electrical signal by using a photoelectric conversion device (PMT or SiPM). The front-end circuit measures the energy and time of the analog electric signal with high precision, and then sends the measurement result to a back-end Data acquisition system (DAQ) for coincidence judgment and selection. And finally, reconstructing a radiation source image by using the time difference of two events of the effective signal on the coincidence line. In addition to the effect of the signal-to-noise ratio of the detector hardware itself, the accuracy of the Time-to-Digital Converter (TDC) also directly affects the image quality.

For time measurement in conventional PET systems, time measurement is typically done with the aid of high precision commercial Application Specific Integrated Circuit (ASIC) TDC chips. The ASIC TDC has the advantages of good stability and high accuracy, but it is costly, complex in circuitry, and not conducive to ultra-multi-channel applications such as PET. In recent years, the TDC based on the FPGA has been developed rapidly due to its advantages of low cost, short development period, high precision, flexible configuration, and the like. The FPGA-TDC based on the carry chain type is realized by utilizing the carry chain resource inherent in the FPGA, has the advantage of low development difficulty, and is popular. The principle is that the arrival time of the input pulse is marked by using the different level states generated by the input pulse signal on each carry unit of the carry chain. The precision of the FPGA-TDC based on the carry chain can generally reach tens of picoseconds due to the carry unit delay and the nonuniformity of the carry unit delay. In order to further improve the measurement accuracy and overcome the limitation of a minimum delay unit, a Wave-Union method is proposed by the Fermi national laboratory in the United states, and the measurement accuracy can be improved under the condition of saving FPGA resources as much as possible. The general principle is that Wave Union launchers are introduced into a carry chain to excite an input pulse signal to generate stable periodic oscillation, and the purpose of improving the measurement precision is achieved by measuring the leading edge time of the signal for multiple times.

How to apply the Wave-unity FPGA-TDC device to PET becomes a problem to be solved by the scheme.

Disclosure of Invention

In order to solve the above problems, an object of the present invention is to provide a Wave-unity type TDC apparatus and a measurement method for PET.

In order to achieve the purpose, the technical scheme of the invention is as follows:

a PET Wave-unity type TDC device comprises an input end D trigger, a multiplexer, a NOT gate, a carry chain, a Wave-unity Launcher, a DFF array, a pipeline encoder, nonlinear correction logic, fine time calculation logic, a coarse counter and a data buffer;

the input end D trigger is used for latching and clearing an input pulse signal;

the CARRY chain device is formed by connecting CARRY4 hard cores in the FPGA in series, input pulse signals are transmitted in the chain in an incremental manner after entering, and the pulse arrival time is marked by using different tap level states on the chain;

the DFF array is composed of D triggers and used for latching the level on a carry chain tap and generating a thermometer code;

the Wave-unity Launcher utilizes one tap of a carry chain and feeds back the tap to the input end of a multiplexer through an NOT gate to generate periodic oscillation pulse;

the Pipeline encoder is used for converting thermometer codes into binary coded fine time information;

the fine time calculation logic is used for normalizing fine time information obtained by multiple times of oscillation and summing or averaging to obtain a final fine time result;

the non-linear correction logic is to correct a final fine time result;

the coarse counter is an FPGA internal counter soft core driven by a sampling clock and is used for obtaining coarse time information;

the data buffer is used for buffering the obtained time data and communicating with other logics of the system.

Further, the carry chain delay Δ chain should be greater than the sampling clock period Tclk.

Further, the CARRY4 includes four taps, where the level states of two taps, which have consistent and minimal delays to the DFF array, are used to mark the pulse arrival time.

Further, the Pipeline encoder consists of a one-hot code encoder and a binary encoder; the one-hot code encoder obtains one-hot codes by detecting thermometer codes generated by DFF array latch; the binary encoder comprises a basic tree structure and an output tree structure, wherein the basic tree structure utilizes an OR gate to build a layer tree structure, the output tree structure extracts the middle level of the basic tree structure, and finally the one-hot code is converted into binary encoded fine time information.

Furthermore, a D flip-flop array is added after each layer of gate of the basic tree structure and the output tree structure.

A measuring method of a Wave-unity type PET TDC device comprises the following steps:

s01, inputting a pulse signal to be detected, enabling the pulse signal to be detected to enter a carry chain after passing through an input end D trigger and a multi-path selector, selecting 1 by the multi-path selector when the pulse signal is transmitted to a kth tap, and pulling down an input signal of the carry chain; when the low level reaches the kth tap again, the multiplexer selects 0, the carry chain input signal is pulled high, so that periodic oscillation pulses are generated, and the oscillation period Tosc of the Wave-unity Launcher is greater than the carry chain delay delta chain; after an input pulse signal enters, the input pulse signal is transmitted in a chain in an incremental mode, the pulse arrival time is marked by using different tap level states on the chain, and the carry chain delay delta chain is larger than the sampling clock period Tclk;

s02, the DFF array latches the level on the carry chain tap and generates thermometer code;

s03, the one-hot code encoder obtains the one-hot code by detecting the thermometer code, and the binary encoder converts the one-hot code into binary encoded fine time information;

s04, the fine time calculation logic is to normalize the fine time information obtained by multiple times of oscillation, and sum and average to obtain the final fine time result, the normalization is performed by a recursive operation formulaImplementation of, wherein t_iThe fine time of each measurement is represented, and the alpha type means that two pulse signals to be measured are respectively detected in two clock cycles; the beta type means that the same pulse signal to be detected is detected in two clock cycles, and the difference is one clock cycle; for the oscillation period Tosc, continuously measuring 0 pulse signals to be measured; for the beta and gamma classes, each measurement cycle is compensated for one cycle in the measurement results, for the alpha class, oneThe above recursive operation can be normalized to fine time information relative to the first event; adding or averaging the fine time information measured each time to obtain a final fine time result;

s05, the nonlinear correction logic is used for correcting the final fine time result;

s06, the coarse counter uses the FPGA internal counter soft core driven by the sampling clock to obtain the coarse time information;

and S07, combining the coarse time information and the corrected fine time result to complete time measurement in a dynamic range.

Further, in step S04, the fine time information for each measurement is averaged, and the average calculation formula is:

，

where M is the total number of measurements, t (M,1) is the fine time result relative to the first measurement, t (M) is the fine time result of this measurement, and n (M) is the measurement period function that needs compensation for both beta and gamma type events.

Further, the oscillation period Tosc is determined by the following formula:

wherein, Δ tap is the carry chain tap average unit delay, Δ MUX is the multiplexer delay, Δ INV is the not gate delay, and Δ ROUT is the routing delay; the determination of the fractional delay comprises the steps of: and fixing the positions of the NOT gate and the multiplexer in the FPGA by using a development tool, carrying out delay time sequence constraint and routing fixation on the delta ROUT, and finally estimating the delay of each part according to an imitation result after tool wiring.

Further, the step S03 further includes the following steps: carrying out NAND operation on each bit of the thermometer code and adjacent high four bits of the thermometer code, only marking the position where the highest jump of '01' occurs, and respectively carrying out AND operation on the highest three bits and the lowest bit, wherein the Boolean operation formula is as follows:

。

further, the step S05 includes: the nonlinear correction logic utilizes random pulse signals generated by background noise of a detector, counts the occurrence frequency of each fine time when the system is started, calculates the delay uniformity, obtains nonlinear correction parameters by utilizing a bin-by-bin method, stores the nonlinear correction parameters in a Random Access Memory (RAM) in a Field Programmable Gate Array (FPGA) as a lookup table, and reads the correction values in the lookup table according to the final measured fine time result as an address to obtain the corrected fine time value.

The invention has the advantages that:

1. based on carry chain type FPGA-TDC, a Wave Union Launcher is introduced to realize multiple time measurement and improve the measurement precision;

2. the tree-shaped encoder with the pipeline function can accurately, quickly and stably convert binary codes generated by Wave-unity into time codes;

3. the functions of online nonlinear correction, clock frequency calibration and the like are realized by utilizing the background noise of the detector;

4. instantiating in Xilinx Kintex 7 series FPGA and performing electronic preliminary test, wherein test results show that the measurement precision of the FPGA-TDC is better than 15 ps RMS, and the dead time is less than 20 ns; in addition, the invention also tests on an actual PET detector, and the test result shows that the TOF measurement time resolution is better than 320 ps FMHW, thereby meeting the performance requirement.

Drawings

FIG. 1 is a block diagram of the overall structure of the Wave-Union type TDC device in an embodiment;

FIG. 2 is a schematic structural diagram of a Pipeline encoder in the embodiment;

FIG. 3 is a table of simulated results after delay of taps of different CARRY4 to adjacent D flip-flops in an embodiment;

FIG. 4 shows the single-channel cycle test results of the FPGA-TDC in the embodiment;

FIG. 5 shows the coincidence time resolution test results of the PET detector in the embodiment.

Detailed Description

The present invention will be described in further detail with reference to examples.

The embodiment provides a PET Wave-unity TDC device, as shown in fig. 1, including an input end D flip-flop, a multiplexer, a not gate, a carry chain, a Wave-unity Launcher, a DFF array, a pipeline encoder, a non-linear correction logic, a fine time calculation logic, a coarse counter, and a data buffer.

The input end D flip-flop is used for inputting, buffering and clearing pulse signals.

The CARRY chain is formed by connecting CARRY4 hard cores in the FPGA in series, input pulse signals are transmitted in the chain in an incremental mode after entering, and the pulse arrival time is marked by using different tap level states on the chain. In order to cover the complete measurement range with the fine count value, the carry chain delay Δ chain should be longer than the sampling clock period (denoted as Tclk), the sampling clock is used to latch the levels on the carry units of the carry chain, if the sampling period is longer than the length of the carry chain, some level states cannot be captured, and only if the period is shorter than the delay time of the signal on the carry chain, the number of delay units passed by the input signal on the carry chain can be known. In this embodiment, to reduce the measurement dead time, the sampling clock period is set to 4 ns. Through testing, one CARRY4 hardmac delay is approximately around 40ps, so the present invention commonizes 110 CARRY4 hardmacs, each CARRY4 includes 4 taps, the present embodiment uses two of which as flags, totaling 220 tap levels.

The DFF array consists of 220D flip-flops to latch the levels on the carry chain taps. To ensure uniformity of the carry chain, the delay of the carry chain taps to the DFF needs to be uniform and as small as possible. Using the Xilinx development tool Vivado, post-wiring timing simulations were performed on the delays of the eight nearest-neighbor D-flip-flops to the same Slice for four taps in a CARRY4 hardmac, with the results shown in fig. 3, where the delay units are ps. According to simulation results, tap 0 and tap 2 are connected to DFF7 and DFF3 respectively, so that small and similar delay can be obtained, and the purpose of improving the uniformity of the carry chain is achieved (in the table, DFF2 corresponding to tap 1 also has small delay, but because the delay is large when the carry chain is crossed, the whole chain is more uniform by selecting taps 0 and 1 or 1 and 2).

The Wave-unity Launcher uses one of the taps of the carry chain to feed back to the input of the multiplexer through the not gate. When the pulse signal passes to the kth tap, the multiplexer selects 1 and the carry chain input signal is pulled low. When the low level reaches the kth tap again, the multiplexer selects 0 and the carry chain input signal is pulled high, thus generating a periodic oscillation pulse. After a specified number of pulses are generated, the multiplexer is set to 0 and clears the D flip-flop at the input end to wait for the next pulse signal to arrive. To ensure that only one signal front appears at most on a carry chain, the oscillation period Tosc of the Wave-unity Launcher should be larger than the carry chain delay Δ chain. Tosc is determined approximately by the following formula:

wherein, Δ tap is the carry chain tap average unit delay, Δ MUX is the multiplexer delay, Δ INV is the not gate delay, and Δ ROUT is the routing delay; for the FPGA chip, the delta tap, the delta INV and the delta MUX are basically constant values, and the biggest influence is delta ROUT. In order to ensure that the Tosc meets the design requirements, a development tool is used for fixing the positions of the inverter and the multiplexer in the FPGA, and carrying out delay timing constraint and routing fixation on the delta ROUT, and finally, the delay of each part is estimated according to the simulation result after the tool is wired. In the present invention, K is 25.

The Pipeline encoder consists of a one-hot code encoder and a binary encoder. The one-hot code encoder is a device for detecting the location of a "01" transition in the thermometer code (e.g., "00001111") generated by the DFF array latch. Each bit of the thermometer code respectively performs NAND operation with adjacent high bits, the NAND operation condition is met only at the jump position of '01', finally the one-hot code (such as '00001000') is obtained, and the one-hot code (such as '00001000') is finally obtained. In order to reduce the abnormal jump (such as "00010111") caused by the metastable state of the DFF latch, each bit of the thermometer code and the adjacent upper four bits are utilized to carry out NAND operation, only the position where the highest jump of "01" appears is marked, and for the highest three bits, the highest three bits are respectively AND-or-operated with the lowest bit, and the Boolean operation formula is as follows:

。

the binary encoder is composed of two parts, one part is a basic tree structure, and the other part is an output tree structure. The basic tree structure utilizes an OR gate to build a layer tree structure, and the output tree structure extracts the intermediate level of the basic tree structure and finally converts the one-hot code into a binary code; in order to ensure the coding stability, a D flip-flop array is added behind each layer of gate of the basic tree structure and the output tree structure, so that a Pipeline structure is realized.

The fine time calculation logic normalizes the fine time information obtained by multiple oscillations, and sums or averages the fine time information to obtain a final fine time result.

Nonlinear correction logic is used to correct the final fine time result.

The coarse time counter is an FPGA internal counter soft core driven by a sampling clock and is used for obtaining coarse time information (the minimum time measurement interval bin size of the coarse time is 4 ns). The combination of the coarse time information and the final fine time information can simultaneously ensure the measurement precision and the measurement dynamic range. In this embodiment, the coarse time counter is 40 bits, and the dynamic range of measurement exceeds 3 hours.

The data buffer is used for buffering the obtained time data and communicating with other logics of the system.

The measuring method of the PET Wave-unity type TDC device comprises the following steps:

s01, inputting a pulse signal to be detected, enabling the pulse signal to be detected to enter a carry chain after passing through an input end D trigger and a multi-path selector, selecting 1 by the multi-path selector when the pulse signal is transmitted to a kth tap, and pulling down an input signal of the carry chain; when the low level reaches the kth tap again, the multiplexer selects 0, the carry chain input signal is pulled high, so that periodic oscillation pulses are generated, and the oscillation period Tosc of the Wave-unity Launcher is greater than the carry chain delay delta chain; after an input pulse signal enters, the input pulse signal is transmitted in a chain in an incremental mode, the pulse arrival time is marked by using different tap level states on the chain, and the carry chain delay delta chain is larger than the sampling clock period Tclk;

s02, the DFF array latches the level on the carry chain tap and generates thermometer code;

s03, the one-hot code encoder obtains the one-hot code by detecting the thermometer code, and the binary encoder converts the one-hot code into binary encoded fine time information;

s04, the fine time calculation logic is to normalize the fine time information obtained by multiple oscillations, and sum and average the information to obtain the final fine time result, as mentioned above, this embodiment satisfies the condition Tclk<Δchain<Tosc, so only 0 or 1 leading edge can be detected in the carry chain per clock cycle. The normalization is by a recursive operationImplementation of, wherein t_iThe fine time of each measurement is represented, and the alpha type means that two pulse signals to be measured are respectively detected in two clock cycles; the beta type means that the same pulse signal to be detected is detected in two clock cycles, and the difference is one clock cycle; for the oscillation period Tosc, continuously measuring 0 pulse signals to be measured; according to the formula, the fine count obtained by each measurement is compared with the previous measurement value, which type the fine count belongs to is judged, for beta and gamma types, each measurement period compensates one period in the measurement result, and for alpha type, the fine time information relative to the first event can be normalized through the recursive operation of the formula; averaging the fine time information measured each time to obtain a final fine time result, wherein the averaging calculation formula is as follows:

，

wherein, M is the total number of measurements, t (M,1) is the fine time result relative to the first measurement, t (M) is the fine time result of this measurement, n (M) is the measurement period function to be compensated for the beta and gamma type events (if the leading edge of an oscillation signal is detected in a period, n (M) is 0, no compensation is needed, if the leading edge of the oscillation signal is not detected, but is detected in another period, n (M) is 1, and a period needs to be compensated in the middle);

s05, the nonlinear correction logic utilizes random pulse signals generated by background noise of a detector, counts the occurrence frequency of each fine time and calculates the delay uniformity when the system is started, a bin-by-bin method is utilized to obtain nonlinear correction parameters, the nonlinear correction parameters are used as a lookup table to be stored in a Random Access Memory (RAM) inside a Field Programmable Gate Array (FPGA), and the actual pulse signals to be detected are used as addresses to read the correction values in the lookup table according to the final measured fine time result to obtain corrected fine time values; the nonlinear correction parameters are used as a Look-up table (LUT) and stored in a Random Access Memory (RAM) in the FPGA, the actual signal to be detected is used as an address to read the correction value in the LUT according to the measured fine time information to obtain a corrected fine time value, the number of each fine time value can be counted by measuring a large number of random pulse signals according to a code density method, and the number of the numbers is in direct proportion to a tap delay value;

for example, there are two random access memories RAM1 and RAM2 in FPGA, and RMA1 reads the content w of address i_iThe nonlinear correction coefficient D is calculated in turn by using the following formula_i：

，

Writing the calculated nonlinear correction coefficient as content into RMA2 as a correction lookup table LUT, inputting the actual pulse signal to be measured into a carry chain for fine time measurement, and inquiring the content in RMA2 by using the final fine time measurement result as an address to obtain a corrected fine time value;

in addition, in order to facilitate calculation such as coincidence of the rear-end DAQ, the design sets the total number of statistical random pulses to be the power N of 2, and finally normalizes the fine counting format through shift operation;

s06, the coarse counter uses the FPGA internal counter soft core driven by the sampling clock to obtain the coarse time information;

and S07, combining the coarse time information and the corrected fine time result to complete time measurement in a dynamic range.

The TDC was instantiated on a Xilinx K7 FPGA and subjected to preliminary verification testing. The signal source generates a series of periodic pulses that are input to the TDC. After 3 passes of nonlinear correction and Wave Union, the fine time is normalized to 10 bits with a binsize of 3.9 ps and a resolution of about 14.9 ps RMS, as shown in fig. 4.

In addition, a detector coincidence test is carried out by utilizing two Lyso crystals (the length of the crystals is 15 cm), a Na22 radioactive source is placed in the middle of the small module, and energy, position and time information measured by the system are collected. In order to screen the effective signal, an energy window (425 keV to 650 keV) of the screening energy is set, and the time resolution test result is shown in FIG. 5. Under the condition of a Na22 point source, the two small modules meet the requirement of TOF imaging, and the time resolution is better than 320 ps FWHM.

The above-mentioned embodiments are merely illustrative of the inventive concept and are not intended to limit the scope of the invention, which is defined by the claims and the insubstantial modifications of the inventive concept can be made without departing from the scope of the invention.