Modeling and simulation methods for total dose effect and instantaneous dose rate effect of microprocessor
By combining behavioral-level models and photocurrent equations, the problem of high modeling complexity of total dose effect and instantaneous dose rate effect in microprocessors is solved, achieving efficient simulation results and making it suitable for complex devices at the microprocessor level.
Patent Information
- Application Number
- CN202310332208.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-03-30
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2043-03-30
AI Technical Summary
Existing modeling methods for total dose effect and instantaneous dose rate effect in microprocessors are mainly based on physical modeling, which leads to high modeling complexity, slow simulation speed, and the inability of independent simulation technology to meet the complex device requirements of mixed-signal circuits.
The total dose effect and instantaneous dose rate effect of the microprocessor were constructed using a behavioral-level model. A measurement board was designed based on the IBIS model measurement principle to measure the IBIS model data of the microprocessor, constructing a FuncIBS model. The model was then simulated using the photocurrent equation model, and the model was verified using SystemVision software.
The modeling complexity of total dose effect and instantaneous dose rate effect of microprocessors is reduced, the simulation efficiency is improved, and accurate simulation of complex devices at the microprocessor level is achieved, which conforms to the actual working characteristics.
Smart Images

Figure CN116362173B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of computer simulation technology, and in particular relates to a modeling and simulation method for the total dose and instantaneous dose rate effect of a microprocessor, which can be used in a simulation model system for the irradiation effect of digital devices. Background Technology
[0002] Currently, with the rapid development of semiconductor technology, various analog and digital integrated circuits are widely used in aerospace, nuclear industry, and particle physics. These circuits operate in environments exposed to electromagnetic and high-energy particle radiation, and are subject to various radiation effects such as total dose radiation, single-event radiation, and transient effects, severely testing their operational reliability and lifespan. Among these, the total dose radiation effect has received particular attention and research. Therefore, constructing a model of the total dose radiation effect on devices, simulating the constructed model, and simulating bit-flipping faults caused by the total dose radiation effect—that is, the fault injection method for bit-flipping—has become crucial.
[0003] In recent years, with the development of computer simulation technology, the simulation of radiation effects in electronic systems has provided strong support for system radiation resistance design and theoretical research, shortening the R&D cycle and reducing development costs. Currently, modeling research on the total dose effect and instantaneous dose rate effect of microprocessors mainly falls into two categories: transistor-level models (SPICE) based on underlying physics and behavioral models (IBIS) based on functionality. Transistor-level models offer high simulation accuracy and wide applicability, but they have high computational overhead, limited circuit scale for simulation, and, due to the intellectual property involved in the physical models, semiconductor manufacturers are often unwilling to provide corresponding models or encrypt them, making them difficult to use for modeling total dose effects and instantaneous dose rate effects. Behavioral models, on the other hand, do not involve underlying physics, are simple, have good versatility, and high simulation efficiency.
[0004] Patent document with application number 2016108244747 discloses "a method for modeling total dose effect of integrated circuits based on IBIS". This method is based on the IBIS model of integrated circuits to model the total dose effect of devices, and adopts a behavioral-level model.
[0005] Patent document with application number CN113341761A discloses "a method and system for modeling the total dose effect of CMOS digital integrated circuits". This method is for CMOS digital integrated circuits and models the total dose effect of digital integrated circuits, using the same behavioral model.
[0006] Although both of the above methods use behavioral-level models, they are for simple devices, such as gate circuits, and do not go up to the level of building behavioral-level models for complex devices at the microprocessor level.
[0007] While independent simulation of digital circuits and independent simulation of analog circuits are still used in current simulation technology, they can no longer meet the needs of most designs in the era of mixed-signal circuits. Mixed-signal simulation technology solves this problem well. It is used to simulate the total dose effect and instantaneous dose rate effect models of microprocessors to meet more application requirements.
[0008] The patent document with application number CN113341761A proposes a "method and system for modeling the total dose effect of CMOS digital integrated circuits". After modeling the total dose effect of digital integrated circuits, the method uses mixed signal simulation technology to simulate the model.
[0009] Patent document with application number CN106503285B discloses "a circuit irradiation effect modeling method based on VHDL-AMS". This method is based on VHDL-AMS to model the irradiation effect of the circuit and also uses mixed signal simulation technology to simulate the model.
[0010] Although both of the above methods use mixed-signal simulation technology to simulate the model, they only simulate the irradiation effect model of simple devices such as gate circuits, and do not simulate the irradiation effect model of complex devices at the microprocessor level, let alone the irradiation effect model combined with the microprocessor IP core.
[0011] Research on fault injection methods for bit flips mainly falls into two categories: those based on low-level physics and those based on high-level behavioral simulation models. Low-level physics-based fault injection methods induce state bit changes in sensitive nodes by adding current sources to the circuit, offering wide applicability, but are highly complex and time-consuming for large-scale integrated circuits with millions of gates or more. High-level behavioral simulation model-based fault injection methods, on the other hand, can directly inject bit flips into the machine code of the ROM area, offering a simpler injection method with lower simulation difficulty and higher efficiency.
[0012] In summary, existing modeling methods for total dose effect and instantaneous dose rate effect of microprocessors can only be implemented based on physical underlying modeling, which is complex, slow in simulation, and independent simulation technology can no longer meet the needs of use. Summary of the Invention
[0013] The purpose of this invention is to address the shortcomings of the prior art by providing a modeling and simulation method for the total dose effect and instantaneous dose rate effect of a microprocessor, thereby reducing the complexity of constructing physical models of the total dose effect and instantaneous dose rate effect of a microprocessor and improving simulation efficiency.
[0014] The technical solution of the present invention includes a microprocessor-based total dose effect modeling and simulation method and an instantaneous dose rate effect modeling and simulation method, the implementation steps of which are as follows:
[0015] 1. A method for modeling and simulating the total dose effect of a microprocessor, characterized by comprising the following steps:
[0016] (1) Based on the IBIS model measurement principle and the internal circuit structure of the microprocessor, design the IBIS model measurement board of the microprocessor;
[0017] (2) Measure the IBIS model data of the microprocessor using a measurement circuit board;
[0018] (3) Construct the microprocessor total dose-effect model FuncIBS based on the microprocessor's IBIS model data:
[0019] (3a) Match the VI characteristic curves of the ground clamping transistor and the VI characteristic curves of the power clamping transistor in the IBIS model data with the input ports, and match the output characteristic curves of the pull-up transistor, the pull-down transistor, and the VT characteristic curves of the output buffer in the IBIS model data with the output ports.
[0020] (3b) Connect the input terminal of the microprocessor to its functional area via an analog-to-digital converter (A / D), and connect the functional area of the microprocessor to its output terminal via a digital-to-analog converter (D / A) to form the microprocessor total dose-effect model FuncIBS:
[0021] (4) The obtained microprocessor total dose effect model FuncIBS was simulated. The simulation results of the rising and falling edges of the model at total doses of 0 Krad, 10 Krad, 20 Krad, and 50 Krad were compared with the measured results, and the relative deviation α between the simulation results and the measured results was calculated. D ;
[0022] (5) Determine the obtained deviation α D Does it meet the required technical specifications?
[0023] If so, it indicates that the established simulation model is more accurate than the measurement results. Using this model, the characteristics of the microprocessor can be described more accurately, so as to obtain the required data better.
[0024] Otherwise, the model cannot accurately describe the characteristics of the microprocessor.
[0025] 2. A modeling and simulation method for the instantaneous dose rate effect of a microprocessor, characterized by comprising the following steps:
[0026] 1) Based on the IBIS model measurement principle and the internal circuit structure of the microprocessor, design the IBIS model measurement board for the microprocessor;
[0027] 2) Measure the IBIS model data of the microprocessor using a measurement circuit board;
[0028] 3) Based on the microprocessor's IBIS model data, construct the microprocessor instantaneous dose rate effect model FuncIBS:
[0029] 3a) Match the VI characteristic curves of the ground clamp transistor and the power clamp transistor in the IBIS model data with the input ports, and match the output characteristic curves of the pull-up transistor, pull-down transistor, and VT characteristic curves of the output buffer with the output ports.
[0030] 3b) Connect the input terminal of the microprocessor to its functional area via an analog-to-digital converter (A / D), and connect the microprocessor's functional area to its output terminal via a digital-to-analog converter (D / A) to form the microprocessor total dose-effect model FuncIBS:
[0031] 4) Obtain photocurrent data of the microprocessor through instantaneous dose rate experiments, and establish a microprocessor instantaneous dose rate effect model:
[0032] 4a) Based on the instantaneous dose rate experiment, the photocurrent data of the microprocessor is obtained, and the parameters of the experimentally obtained photocurrent are obtained according to the meaning of the four-parameter Weibull function;
[0033] 4b) Construct fitting parametric equations based on the influence of dose rate on photocurrent equation parameters, namely, equation y for dose rate and time to peak, equation z for dose rate and peak value, and equation f for dose rate and pulse width.
[0034] y = -2.3584e -28 *x 3 +6.02866e -20 *x 2 -5.265e -12 *x+1.4534e -4
[0035] z = 1.1876e -3 1*x 4 -2.595e -23 *x 3 +1.9191e -15 *x 2 -5.4034e -8 *x+0.52514
[0036] f = -4.6858e -35 *x 4 +1.0036e -26 *x 3 -7.1535e -19 *x 2 +1.881e -11 *x-8.7749e -5
[0037] Where y represents the time to reach the peak value, z represents the peak value, f represents the pulse width, and x represents the dose rate;
[0038] 4c) Describe the fitted parametric equations using VHDL-AMS language, construct a photocurrent model, and add this model to the output of the microprocessor FuncIBS model to obtain the microprocessor instantaneous dose rate effect model, which includes the microprocessor's input buffer, functional area, and output buffer.
[0039] 5) Simulate the microprocessor instantaneous dose rate effect model:
[0040] 5a) Load the design file of the microprocessor instantaneous dose rate effect model into the SystemVision simulation software, and set the dose rate parameters to 8.85e6Gy / s, 1.4e7Gy / s, 1.9e7Gy / s, 4.35e7Gy / s, 7.4e7Gy / s, 9.65e7Gy / s, and 5.55e7Gy / s;
[0041] 5b) Observe whether the square wave output and photocurrent output of the individual microprocessor of the model are correct at the dose rate of 5a):
[0042] If correct, then apply photocurrent to the port of the output square wave;
[0043] Otherwise, return to 4b).
[0044] Compared with the prior art, the present invention has the following advantages:
[0045] This invention addresses the problem of high complexity and low simulation efficiency of microprocessor physical models by constructing a behavioral model of the microprocessor in the total dose effect modeling and simulation method for microprocessors. This elevates the behavioral modeling of simple devices such as gate circuits to complex devices at the microprocessor level.
[0046] This invention solves the problem of difficult photocurrent simulation of microprocessors under instantaneous dose rate effects by adding a fitted photocurrent equation model to the microprocessor behavioral model based on experimental data in the microprocessor instantaneous dose rate effect modeling and simulation method, and realizes behavioral modeling of the instantaneous dose rate effect of microprocessors. Attached Figure Description
[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0048] Figure 1 This is a flowchart of the microprocessor total dose effect modeling implementation of the present invention;
[0049] Figure 2 This is a diagram of the existing IBIS model structure;
[0050] Figure 3 yes Figure 1 Internal structure diagram of the microprocessor selected in the process;
[0051] Figure 4 yes Figure 1 A schematic diagram of the principle of constructing an IBIS model for measurement;
[0052] Figure 5 yes Figure 1 The structure diagram of the total dose effect model of the microprocessor constructed in the middle;
[0053] Figure 6 yes Figure 1 Comparison of simulation and test results of rising edge waveform of microprocessor at 0 krad total dose;
[0054] Figure 7 yes Figure 1 Comparison of simulated and test results of falling edge waveform of microprocessor at 0 krad total dose;
[0055] Figure 8 yes Figure 1 Rising edge waveforms of the total dose effect model of a microprocessor under different total doses;
[0056] Figure 9 yes Figure 1 Falling edge waveforms of the total dose effect model of a microprocessor under different total doses;
[0057] Figure 10 This is a flowchart of the microprocessor instantaneous dose rate effect modeling implementation of the present invention;
[0058] Figure 11 yes Figure 10 Simulation waveforms of square wave and photocurrent under different dose rates in the microprocessor instantaneous dose rate effect model;
[0059] Figure 12 yes Figure 10Simulation waveforms of square wave superimposed photocurrent at different dose rates in the microprocessor instantaneous dose rate effect model;
[0060] Figure 13 yes Figure 10 The structure diagram of the microprocessor instantaneous dose rate effect model constructed in the model. Detailed Implementation
[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0062] To address the problems existing in the prior art, this invention provides a microprocessor-based total dose effect modeling and simulation method and an instantaneous dose rate effect modeling and simulation method. The following describes these two examples of the invention in detail with reference to the accompanying drawings.
[0063] Example 1: Microprocessor-based total dose effect modeling and simulation method.
[0064] Reference Figure 1 The implementation steps for this example are as follows:
[0065] Step 1: Based on the IBIS model measurement principle and the internal circuit structure of the microprocessor, design the IBIS model measurement board for the microprocessor.
[0066] This example uses the open-source MC8051 IP core provided by Oregano Systems as the microprocessor, and its circuit structure is as follows: Figure 3 The open-source MC8051 IP core adopts a fully synchronous design, with an instruction set fully compatible with the standard 8051 microcontroller. It has 256 bytes of internal RAM and can be expanded to a maximum of 64 KB of ROM and 64 KB of RAM.
[0067] Reference Figure 2 The IBIS model structure, with the left side of the dashed line representing the input model structure and the right side representing the output model structure, can be defined as a simple set of functions to generate the IBIS model.
[0068] This step designs a microprocessor-based IBIS model measurement board based on the IBIS model measurement principle and the microprocessor's circuit structure. The board includes an input buffer VI characteristic measurement unit, an output buffer VI characteristic measurement unit, and an output terminal VT characteristic measurement unit. Its schematic diagram is shown below. Figure 4 As shown, where:
[0069] The input buffer VI characteristic measurement unit includes a Power Clamp measurement circuit for the VI curve of the input buffer power clamp diode and a GND Clamp measurement circuit for the VI curve of the input buffer ground clamp diode. Both measurement circuits include a first controllable voltage source and a first original device. The first controllable voltage source is used to power the first original device and provide a scan voltage. The positive terminal of the scan voltage of the Power Clamp measurement circuit is connected to the input pin under test of the first original device, and the negative terminal is connected to the power supply pin VCC of the first original device. The positive terminal of the scan voltage of the GND Clamp measurement circuit is connected to the ground pin GND of the first original device, and the negative terminal is connected to the input pin under test of the first original device.
[0070] The output buffer VI characteristic measurement unit includes a pull-up measurement circuit for the pull-up transistor and a pull-down measurement circuit for the VI characteristic curve of the pull-down transistor. Both measurement circuits include a second controllable voltage source and a second original device. The second controllable voltage source is used to power the second original device and provide a scan voltage. The positive terminal of the scan voltage of the pull-up measurement circuit is connected to the power supply pin VCC of the second original device, and the negative terminal is connected to the output pin under test of the second original device. The positive terminal of the scan voltage of the pull-down measurement circuit is connected to the output pin under test of the second original device, and the negative terminal is connected to the ground pin GND of the second original device.
[0071] The output terminal VT characteristic measurement unit includes an output terminal rising waveform measurement circuit and a falling waveform measurement circuit. Both measurement circuits include a controllable voltage source, an oscilloscope, and a raw device. The controllable voltage source is used to power the raw device, and the oscilloscope is used to measure the VT data across the load resistor of the raw device.
[0072] Step 2: Measure the IBIS model data of the microprocessor using a measurement circuit board.
[0073] To measure the VI characteristic curve of the power clamp diode, the original device is connected to the Power Clamp circuit. First, the original device is powered by the first channel of the controllable voltage source. The positive terminal of the second channel of the controllable voltage source is connected to the input pin to be tested of the original device, and the negative terminal is connected to the power supply pin VCC of the original device. Then, the VI characteristic measurement module in the control unit applies a scanning voltage to the second channel of the controllable voltage source, so that the power clamp diode in the input buffer of the device is turned on and the ground clamp diode is turned off. The corresponding voltage and current values are recorded at this time. The obtained data is the VI characteristic curve of the power clamp diode, PowerClamp.
[0074] The measurement of the VI curve of the ground clamping diode is performed by connecting the original device to the GND Clamp measurement circuit. First, the original device is powered by the first channel of the controllable voltage source, and the positive terminal of the second channel of the controllable voltage source is connected to the device's ground pin (GND), and the negative terminal is connected to the input pin to be tested. Then, the VI characteristic measurement module in the control unit applies a scanning voltage to the second channel of the controllable voltage source, causing the input buffer power clamping diode to be cut off and the ground clamping diode to be turned on. The corresponding voltage and current values are recorded at this time, and the obtained data is the VI characteristic curve GND Clamp of the ground clamping diode.
[0075] To measure the VI characteristic curve of the output buffer pull-up transistor, the original device is connected to the pull-up measurement circuit. First, the original device is powered by the first channel of the controllable voltage source, and the positive terminal of the second channel is connected to the device's power supply pin VCC, and the negative terminal is connected to the output pin to be tested. Then, a scanning voltage is applied to the second channel of the controllable voltage source, which turns on the pull-up transistor and turns off the pull-down transistor, and the corresponding current value is recorded. The recorded current value is then subtracted from the power clamp current value, and the resulting data is the pull-up transistor VI characteristic curve.
[0076] To measure the VI characteristic curve of the output buffer pull-down transistor, the original device is connected to the Pulldown measurement circuit. First, the original device is powered by the first channel of the controllable voltage source. The positive terminal of the second channel is connected to the output pin to be tested, and the negative terminal is connected to the ground pin GND. Then, the VI characteristic measurement module in the control unit applies a scanning voltage to the second channel of the controllable voltage source, causing the pull-up transistor to be cut off and the pull-down transistor to be turned on, and the corresponding current value is recorded. The recorded current value is then subtracted from the GND Clamp current value, and the resulting data is the pull-up transistor VI characteristic curve Pulldown.
[0077] For measuring the rising edge VT waveform of the output buffer, the original device is connected to the Rising Waveform measurement circuit. First, the original device is powered by the first channel of the controllable voltage source, and load resistors are connected between the output pin under test and the device VCC, and between the output pin under test and the device GND, respectively. Then, a high-level-low-level excitation is applied to the input terminal of the device, and the VT curve across the load resistor is measured by the VT characteristic measurement module in the control unit. The VT curve is displayed and saved by the VT characteristic measurement module. The obtained VT curve data is the Rising Waveform of the output buffer when the load is connected to the power supply and when the load is grounded.
[0078] To measure the VT waveform at the falling edge of the output buffer, the original device is connected to the Falling Waveform measurement circuit. First, the original device is powered by the first channel of a controllable voltage source, and load resistors are connected between the output pin under test and the device's VCC, and between the output pin under test and the device's GND, respectively. Then, a "low-to-high" excitation is applied to the device's input terminal, and the VT curve across the load resistor is measured using the VT characteristic measurement module in the control unit. The VT curve is then displayed and saved by the VT characteristic measurement module. The obtained VT curve data is the Falling Waveform of the output buffer when the load is connected to the power supply and when the load is grounded.
[0079] Step 3: Based on the microprocessor's IBIS model data, construct the microprocessor total dose effect model FuncIBS, which consists of an input port area, an output port area, and a functional area.
[0080] 3.1) Match the measured VI characteristic curves of the ground clamping transistor and the power clamping transistor from the IBIS model data with the input ports.
[0081] 3.2) Match the output characteristic curves of the pull-up transistor, pull-down transistor, and VT characteristic curve of the output buffer in the IBIS model data with the output ports;
[0082] 3.3) The functional area of the FuncIBS model is designed based on the internal principles of the microprocessor, that is, it is implemented by calling the top-level file of the microprocessor;
[0083] 3.4) Connect the input terminal of the microprocessor to its functional area via an analog-to-digital converter (A / D), and connect the microprocessor's functional area to its output terminal via a digital-to-analog converter (D / A) to form the microprocessor's total dose-effect model FuncIBS1, the structure of which is as follows: Figure 5 As shown;
[0084] Step 4: Simulate the obtained microprocessor total dose effect model FuncIBS.
[0085] In the SystemVision simulation software, the design file of the microprocessor total dose effect model FuncIBS was loaded, and the total dose parameters were set to 0 Krad, 10 Krad, 20 Krad, and 50 Krad. The rising and falling edges of the square wave output of the microprocessor total dose effect model FuncIBS under these total dose parameters were extracted to obtain the simulation results.
[0086] 4.1) After irradiation experiments and tests, under the condition of a total dose of 0 krad, the rising edge of the output port of the microprocessor total dose irradiation experiment was obtained as follows: Figure 6 and falling edge Figure 7 ;
[0087] 4.2) Under the conditions of total dose parameters of 0 krad, 10 krad, 20 krad, and 50 krad, the microprocessor total dose effect model established in step 3 is controlled to output a square wave, and the simulation results of the rising edge are as follows: Figure 8 The simulation results for the falling edge are as follows: Figure 9 ;
[0088] from Figure 8 and Figure 9 It can be seen that the total dose values of the two curves are 0krad, 10krad, 20krad, and 50krad from top to bottom. As the total dose increases, the rising edge data decreases and the curve slows down, while the falling edge data increases and the curve slows down. This is consistent with the theoretical result of the IBIS model, which makes the rising and falling edges of the device slower. At the same time, it can be seen that when the total dose is 50krad, the curve has a significant slowing trend, indicating that the port damage of the microprocessor is more severe.
[0089] 4.3) Compare the simulation results in 4.2 with the test results in 4.1. The results are shown in Table 1.
[0090] Table 1
[0091] Total dose (krad) Rising edge deviation Falling edge deviation 0 19.92% 17.74% 10 20.03% 18.07% 20 22.34% 19.38% 50 30.17% 28.62%
[0092] As shown in Table 1, the relative errors of the simulation of the rising edge at total doses of 0 krad, 10 krad, 20 krad, and 50 krad are 19.92% (0 krad), 20.03% (10 krad), 22.34% (20 krad), and 30.17% (50 krad), respectively; and the relative errors of the simulation of the falling edge are 17.74% (0 krad), 18.07% (10 krad), 19.38% (20 krad), and 28.62% (50 krad), respectively. This indicates that the hybrid model can reflect the total dose effect of the microprocessor well.
[0093] Example 2: Microprocessor instantaneous dose rate effect modeling and simulation method.
[0094] Reference Figure 10 The implementation steps for this example are as follows:
[0095] Step 1: Based on the IBIS model measurement principle and the internal circuit structure of the microprocessor, design the IBIS model measurement board for the microprocessor.
[0096] The specific implementation of this step is the same as step 1 in Embodiment 1.
[0097] Step two: Use the measurement circuit board to measure the IBIS model data of the microprocessor.
[0098] The specific implementation of this step is the same as step 2 in Embodiment 1.
[0099] Step 3: Based on the microprocessor's IBIS model data, construct the microprocessor total dose effect model FuncIBS, which consists of an input port area, an output port area, and a functional area.
[0100] The specific implementation of this step is the same as step 3 in Example 1.
[0101] Step 4: Obtain photocurrent data of the microprocessor through dose rate experiments and establish a microprocessor instantaneous dose rate effect model.
[0102] 4.1) Choose the four-parameter Weibull function:
[0103] This example uses a current pulse based on the four-parameter Weibull function because this function is highly flexible in fitting random data and can well describe the photocurrent waveform caused by the instantaneous dose rate. The equation for the four-parameter Weibull function is:
[0104]
[0105] Where t0 represents the offset of the waveform on the time axis, parameter a represents the peak value H of the pulse waveform, parameter b is the time parameter representing the pulse width, and parameter c represents the trend of the pulse waveform. The value ranges of these parameters are: a>0, b>0, c>1, t≥0, respectively.
[0106] 4.2) Based on the dose rate experiment, the photocurrent data of the microprocessor was obtained. According to the meaning of the four-parameter Weibull function, the time, peak value and pulse width of the photocurrent at different dose rates were obtained.
[0107] 4.3) Construct a parametric equation fitting the effect of dose rate on the photocurrent equation parameters based on the photocurrent parameters:
[0108] 4.3.1) The parametric equation fitting the dose rate to the time to peak concentration t0 is established as follows:
[0109] t0 = -2.3584e-28*x 3 +6.2866e-20*x 2 -5.265e-12*x+0.00014534
[0110] Where x is the dose rate, and * represents the multiplication sign;
[0111] 4.3.2) The parametric equation for fitting the dose rate and peak value a is established as follows:
[0112] a = 1.1876e-31*x 4 -2.595e-23*x 3 +1.19191e-15*x 2 -5.4034e-8*x+0.52514;
[0113] 4.3.3) The parametric equation for fitting the dose rate and pulse width b is established as follows:
[0114] b = -4.6858e-35*x 4 +1.0036e-26*x 3 -7.1535e-19*x 2 +1.881e-11*x+8.7749e-5;
[0115] 4.3.4) Based on the parametric equations established in steps 4.3.1) to 4.3.3), construct the photocurrent model using VHDL-AMS language;
[0116] 4.4) The established photocurrent model is added to the output of the microprocessor FuncIBS model to obtain the microprocessor instantaneous dose rate effect model, the structure of which is as follows: Figure 13 As shown.
[0117] Step 5: Simulate the instantaneous dose rate effect model of the microprocessor.
[0118] 5.1) Under the conditions of instantaneous dose rates of 8.85e6Gy / s, 1.4e7Gy / s, 1.9e7Gy / s, 4.35e7Gy / s, 5.55e7Gy / s, 7.4e7Gy / s, and 9.65e7Gy / s, the microprocessor instantaneous dose rate effect model established in step four is controlled to output a square wave, and photocurrent is output at other ports to obtain the output of the square wave and photocurrent, as shown below. Figure 11 As shown, where:
[0119] Figure 11 a is Figure 10 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 8.85e6Gy / s;
[0120] Figure 11 b is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 1.4e7Gy / s;
[0121] Figure 11 c is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 1.9e7Gy / s;
[0122] Figure 11 d is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 4.35e7Gy / s;
[0123] Figure 11 e is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 5.55e7Gy / s;
[0124] Figure 11 f is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 7.4e7Gy / s;
[0125] Figure 11 g is Figure 11 Simulated waveforms of square wave and photocurrent in the microprocessor instantaneous dose rate effect model at a dose rate of 9.65e7Gy / s;
[0126] from Figure 11 As can be seen, the wave and photocurrent outputs are correct at different dose rates.
[0127] 5.2) Under the conditions of instantaneous dose rates of 8.85e6Gy / s, 1.4e7Gy / s, 1.9e7Gy / s, 4.35e7Gy / s, 5.55e7Gy / s, 7.4e7Gy / s, and 9.65e7Gy / s, a photocurrent is applied to the port of the output square wave to obtain an output of square wave superimposed with photocurrent, as shown below. Figure 12 As shown, where:
[0128] Figure 12 a is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 8.85e6Gy / s using the microprocessor instantaneous dose rate effect model;
[0129] Figure 12 b is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 1.4e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0130] Figure 12 c is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 1.9e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0131] Figure 12d is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 4.35e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0132] Figure 12 e is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 5.55e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0133] Figure 12 f is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 7.4e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0134] Figure 12 g is Figure 11 Simulation waveform of square wave superimposed photocurrent at a dose rate of 9.65e7Gy / s using the microprocessor instantaneous dose rate effect model;
[0135] from Figure 12 As can be seen, when the photocurrent model is added, a pulse is superimposed on the square wave output waveform at the time of photocurrent appearance, which is consistent with the working condition of the microprocessor. This indicates that the hybrid model can better reflect the instantaneous dose rate effect of the microprocessor.
[0136] The accuracy of the microprocessor instantaneous dose rate effect model in this invention can be illustrated by the following instantaneous dose rate bit-flip fault injection procedure and fault injection tool.
[0137] Based on the simulation results of step five in Example 2, a fault injection program and tool for instantaneous dose rate bit reversal were developed. Fault injection was performed using the fault injection tool, and the simulation results were analyzed and compared with the simulation results of the microprocessor instantaneous dose rate effect model to verify the accuracy of the model. The implementation is as follows:
[0138] The fault injection process is as follows: select the original file and the fault file, select the fault type and fault injection method, and call the fault injection program to implement fault injection. Fault types include one-bit flip and consecutive two-bit flip in the storage area, and fault injection methods include random injection and fixed-point injection.
[0139] A fault injection tool was generated by writing an instantaneous dose rate bit flip fault injection program using MATLAB. The fault injection tool can realize the entire fault injection process. It can open the hex file compiled by the integrated development environment and convert it into a binary data dua file that can be read by the IP core ROM area. Fault injection is achieved by calling the fault injection program to change the data in the previously generated dua file. After selecting to start injection, the interface will display the location of the fault, as well as the original data and the fault data. Check whether the fault injection is successful. If it is unsuccessful, select a file again to start a new injection.
[0140] To verify the effectiveness of the instantaneous dose rate bit-flip fault injection program and tool, the commercial microprocessor STC89C82 was selected. This device is instruction-compatible with the MC8051 IP core and has a similar architecture. Its P0 port outputs binary data 00010001, P1 port outputs binary data 00100010, P2 port outputs binary data 00110011, and P3 port outputs binary data 01000100. This data was compiled using a compiler to generate a hex file that can be burned into the physical MC8051. This hex file is used to burn code into the ROM area of the physical MC8051. The contents of the hex file are converted into a dua file format for reading into the ROM area of the MC8051 IP core. The fault injection program and tool are then used to inject faults into the code.
[0141] The flip position is represented by coordinates. In a one-digit flip, (x,y) means that the flip is performed at the x-th row and y-th column of the dua file. In a two-digit flip, (x,y) means that the flip is performed at the x-th row and y-th column of the dua file and the next position.
[0142] The simulation results of the flipping system are compared with the actual test results, as shown in Table 2;
[0143] Table 2
[0144]
[0145] As shown in Table 2, the simulation relative errors at ports P0, P1, P2, and P3 are 2.63% (P0), 3.59% (P1), 2.19% (P2), and 4.32% (P3) for the one-bit flip case, respectively; and 4.72% (P0), 3.24% (P1), 3.93% (P2), and 3.90% (P3) for the two-bit flip case, respectively.
[0146] The simulation and physical verification results show that the system error rates are very close in both simulation and physical testing, i.e., within an acceptable range. This indicates that the hybrid model can effectively reflect the bit flips caused by the instantaneous dose rate effect of the microprocessor, and that the fault injection program and fault injection tool can effectively perform the fault injection function, verifying the accuracy of the microprocessor instantaneous dose rate effect model.
Claims
1. A modeling and simulation method for the instantaneous dose rate effect of a microprocessor, characterized in that, Includes the following steps: 1) Based on the IBIS model measurement principle and the internal circuit structure of the microprocessor, design the IBIS model measurement board for the microprocessor; 2) Measure the IBIS model data of the microprocessor using a measurement circuit board; 3) Construct the microprocessor total dose-effect model FuncIBS based on the microprocessor's IBIS model data: 3a) Match the VI characteristic curves of the ground clamp transistor and the power clamp transistor in the IBIS model data with the input ports, and match the output characteristic curves of the pull-up transistor, pull-down transistor, and VT characteristic curves of the output buffer in the IBIS model data with the output ports. 3b) Connect the input terminal of the microprocessor to its functional area via an analog-to-digital converter (A / D), and connect the microprocessor's functional area to its output terminal via a digital-to-analog converter (D / A) to form the microprocessor total dose-effect model FuncIBS: 4) Obtain photocurrent data of the microprocessor through dose rate experiments and establish a microprocessor instantaneous dose rate effect model: 4a) Based on the dose rate experiment, the photocurrent data of the microprocessor is obtained, and the parameters of the experimentally obtained photocurrent are obtained according to the meaning of the four-parameter Weibull function; 4b) Construct a fitting parametric equation based on the influence of dose rate on the photocurrent equation parameters, i.e., the equation relating dose rate and time to peak. Equation of dose rate versus peak value Equation for dose rate versus pulse width ; ; ; ; Where y represents the time to reach the peak value. Indicates peak value. Indicates pulse width. Indicates dose rate; 4c) Describe the fitted parametric equations using VHDL-AMS language, construct a photocurrent model, and add this model to the output of the microprocessor FuncIBS model to obtain the microprocessor instantaneous dose rate effect model, which includes the microprocessor's input buffer, functional area, and output buffer. 5) Simulate the microprocessor instantaneous dose rate effect model: 5a) Load the design file of the microprocessor instantaneous dose rate effect model into the SystemVision simulation software, and set the dose rate parameters to 8.85e6Gy / s, 1.4e7Gy / s, 1.9e7Gy / s, 4.35e7Gy / s, 7.4e7Gy / s, 9.65e7Gy / s, and 5.55e7Gy / s; 5b) Observe whether the square wave output and photocurrent output of the individual microprocessor of the model are correct at the dose rate of 5a): If correct, then apply photocurrent to the port of the output square wave; Otherwise, return 4b).
2. The method as described in claim 1, characterized in that, In step 5b), observe whether the square wave output and photocurrent output of the individual microprocessor of the model are correct at the above dose rate, as follows: 5b1) Overlay and compare the microprocessor output square wave and the input square wave at the dose rate in 5a): If the overlap is higher than 75%, it indicates that the square wave output is correct; Otherwise, it indicates that the square wave output is incorrect; 5b2) Overlay and compare the photocurrent output of the microprocessor at the dose rate in 5a) with the photocurrent image obtained from the dose rate experiment: If the overlap is higher than 75%, it indicates that the photocurrent output is correct; Otherwise, it indicates that the photocurrent output is incorrect.
Citation Information
Patent Citations
A circuit irradiation effect modeling method based on VHDL-AMS
CN106503285B
Complementary metal oxide semiconductor (CMOS) digital integrated circuit total dose effect modeling method and system
CN113341761A
