Modeling method of INP HBT large signal model with Knee-Soft effect

Abstract

The invention provides a modeling method of an INP HBT large signal model with a Knee-Soft effect. The modeling method comprises the following steps: acquiring actual measurement data of an INP HBT device for performing a direct current electrical characteristic test; constructing an INP HBT collector direct current model, and obtaining initial values of Knee-Soft effect related parameters in the INP HBT collector direct current model based on measured data; simulating the INP HBT collector direct current model to obtain simulation data of a collector under different electric signals; embedding the INP HBT collector current direct current model into the large signal model, and establishing the large signal model; the simulation data and the actual measurement data are input into the large signal model, the INP HBT collector current direct current model carries out parameter fitting optimization on the input data, and if the fitting degree of all parameters reaches the standard, a final INP HBT large signal model is obtained; through the INP HBT large signal model, the working state of a device can be accurately reflected, and the accuracy of the model and the accuracy of integrated circuit design are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of radio frequency device modeling, and particularly to a modeling method, a computer-readable medium, and a computer program for an INP HBT large-signal model with a Knee-Soft effect. Background Art

[0002] HICUM is a circuit model designed for BJTs and HBTs. With the continuous progress of semiconductor technology, especially the increasing application of HBTs in high-frequency and high-speed electronic circuits, the need for accurate modeling of HBT devices has become increasingly important. INP HBT is crucial for the development of microwave, millimeter-wave, and even terahertz electronic circuits. Modern circuits require high-performance devices with special electrical characteristics. Key parameters, such as high cut-off frequency (ft) / maximum oscillation frequency (fmax), as well as the output power density and breakdown voltage provided by INP HBT, are very suitable for these requirements. Therefore, INP HBT plays an important role in wireless communication, military radar, and optical communication applications.

[0003] The development of HBT devices will inevitably drive the progress of modeling technology. So far, several models have been proposed for HBTs, namely Gummel-Poon, VBIC, MEXTRAM, HICUM, UCSD, Agilent-HBT, and FBH. In traditional INP HBTs, a soft knee effect can be observed in the IC-VC curve at low collector-emitter (CE) voltages, and this soft knee effect can lead to inaccurate responses of the operating state of InP HBTs. Simulating this effect is crucial for enhancing the large-signal performance in high input power and switching applications. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the present invention provides a modeling method for an INP HBT large-signal model with a Knee-Soft effect, so as to solve the problem that the existing INP HBT model cannot accurately reflect the operating state of InP HBTs.

[0005] In order to achieve the above object, the present invention provides a modeling method for an INP HBT large-signal model with a Knee-Soft effect, and its specific steps include:

[0006] Obtain the measured data of the INP HBT device for direct current electrical characteristic testing;

[0007] Construct a DC model of the INP HBT collector, and based on the measured data, obtain the initial values of the parameters related to the Knee-Soft effect in the DC model of the INP HBT collector; simulate the DC model of the INP HBT collector to obtain the simulation data of the collector under different electrical signals;

[0008] Embed the INP HBT collector current DC model into the large-signal model to establish the large-signal model; input the simulation data and the measured data into the large-signal model, and the INP HBT collector current DC model performs parameter fitting and optimization on the input data. If the fitting degree of all parameters meets the standard, the final INP HBT large-signal model is obtained.

[0009] Preferably, when constructing the INP HBT collector DC model, its influencing terms include: current transfer influence and base junction current influence;

[0010] The expression for the current transfer influence is:

[0011]

[0012] The expression for the base junction current influence is:

[0013]

[0014] Among them, i T is the total transfer current, i Tf is the forward transfer current, i Tr is the reverse transfer current; i jBEi is the internal base current, i jBEp is the external base current; i jBEi is composed of the saturation currents I BEiS and I REiS and the non-ideal coefficients m BEi and m REi ; i jBEp is composed of the saturation currents I BEpS and I REpS and the non-ideal coefficients m BEp and m REp ; C 10 is the original parameter of the HICUM model, Q p,T is the hole charge, v B'E' is the voltage between B and E, V T is the turn-on voltage, v B'C' is the voltage between B and C, qks is the empirical model added for the Knee-Soft effect, Kndc, Vc1, kb are the newly added empirical parameters, Vc is the collector voltage, and Ib is the base current.

[0015] Preferably, the expression for the empirical model qks is:

[0016]

[0017] Among them, qk is the new charge introduced in the collector current, which is used to simulate the soft knee effect caused by the base widening effect during the high-level injection process of the collector current in the INP HBT collector current DC model; Q p0 is the original parameter of HICUM, Kn, Kn1, and Kndc1 are newly added empirical parameters, K v is the correction coefficient related to the physical effect of voltage scaling; C 10 is the original parameter of the HICUM model, V B’E’ is the voltage between B and E, V T is the turn-on voltage, and ick is the critical current with high bias effect.

[0018] Preferably, the initial value obtained by the INP HBT collector DC model based on the influencing terms also includes: the initial value unrelated to the Knee-Soft effect;

[0019] When constructing the INP HBT collector DC model, it is divided into a first stage and a second stage according to the Knee-Soft effect. The INP HBT collector DC model in the first stage includes the Knee-Soft effect, and the INP HBT collector DC model in the second stage does not include the Knee-Soft effect.

[0020] Preferably, when the INP HBT collector DC model fits and optimizes the parameters of the first stage, the parameters unrelated to the Knee-Soft effect are obtained; when the INP HBT collector DC model fits and optimizes the parameters of the second stage, the parameters related to the Knee-Soft effect are obtained.

[0021] Preferably, when the INP HBT collector DC model performs parameter fitting, the parameters unrelated to the Knee-Soft effect are first selected for fitting. If the fitting degree does not meet the standard, the unrelated parameters are adjusted and refitted. If the fitting degree meets the standard, the parameters related to the Knee-Soft effect are fitted;

[0022] If the fitting degree of some parameters related to the Knee-Soft effect does not meet the standard, the related parameters are adjusted and refitted. If the fitting degree of all related parameters does not meet the standard, the unrelated parameters are adjusted and refitted; if the fitting degree of all related parameters meets the standard, the overall parameters are adjusted to obtain the INP HBT large-signal model.

[0023] Preferably, the electrical signal includes the collector voltage V C and the base current I B .

[0024] Preferably, when establishing the large-signal model, first construct the topological structure of the large-signal model in the simulation software, and then perform parameter fitting.

[0025] The present invention also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement a modeling method for an INP HBT large-signal model with Knee-Soft effect.

[0026] The present invention also provides a computer program, which is used to execute a modeling method for an INP HBT large-signal model with Knee-Soft effect when the computer program is executed by a processor.

[0027] The beneficial effects of a modeling method for an INP HBT large-signal model with Knee-Soft effect provided by the present invention compared with the prior art are as follows:

[0028] The present invention builds a DC model of the INP HBT collector current through an improved collector current formula, outputs simulation data to the simulation software IC-CAP, and fits and optimizes the simulation data with the measured data in the simulation software IC-CAP, so that the simulation data curve can accurately describe the change of the measured data curve, determines the optimized relevant parameters and irrelevant parameters, and establishes a final large-signal model including Knee-Soft effect, thereby accurately reflecting the working state of the device. Brief Description of the Drawings

[0029] Figure 1 It is a flowchart of a modeling method for an INP HBT large-signal model with Knee-Soft effect provided by the present invention;

[0030] Figure 2 It is a result graph of the model simulation data and the measured data of IC under different VCs in a modeling method for an INP HBT large-signal model with Knee-Soft effect provided by the present invention. Detailed Embodiments

[0031] The following specific examples illustrate the embodiments of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0032] A modeling method for an INP HBT large-signal model with Knee-Soft effect provided by the present invention, as Figure 1 shown, in order to verify the method proposed by the present invention, the present invention provides a specific embodiment, and the specific steps include:

[0033] S1: Conduct direct current electrical characteristic tests on the INP HBT device to obtain the measured data of I under different V C 、I B ; among them, the measured data includes the I C -V C -V C curve.

[0034] In the embodiment of the present invention, the emitter size of the INP HBT device is 0.45um*6um and the base size is 1.05um*6um. Among them, the steps for conducting direct current electrical characteristic tests on the INP HBT device include:

[0035] Under direct current conditions through a test device, scan the collector voltage V C from 0V to 3V with a step size of 0.05V, and scan the base current I B from 0 to 0.001A with a step size of 0.0001A.

[0036] Among them, the test device includes: a vector network analyzer PNA-X, a semiconductor analyzer B1505, a radio frequency probe, and a cable.

[0037] S2: Construct a direct current model of the INP HBT collector current.

[0038] Based on the measured data, obtain the initial values of the parameters related to the Knee-Soft effect in the direct current model of the INP HBT collector current, and simulate the direct current model of the INP HBT collector current in simulation software to obtain the simulation data of I under different V C 、I B ; among them, the measured data includes the I C -V

[0039] S2.1: Construct a direct current model of the INP HBT collector current.

[0040] The establishment of the direct current model of the INP HBT collector current is the core part of the non-linear equivalent circuit model and is crucial for characterizing the power and gain of the device. When constructing the direct current model of the INP HBT collector current, there are two influencing terms, namely the transfer current and the base junction current. Among them, the expression of the influence of the transfer current in the direct current model of the INP HBT collector current is:

[0041]

[0042]

[0043] Among them, i T represents the total transfer current, i TfRefers to the forward current, mainly representing the current where carriers are injected from the emitter region into the base region and reach the collector region after diffusion in the base region. i Tr Refers to the reverse current, representing the situation of carriers being injected from the collector region into the base region in the reverse direction and the current diffusion within the base region. i jBEi Is the internal base current, i jBEp Is the external base current. i jBEi Consists of the saturation current I BEiS And I REiS As well as the non-ideal coefficient m BEi And m REi Constructed. i jBEp Consists of the saturation current I BEpS And I REpS As well as the non-ideal coefficient m BEp And m REp Composed. C 10 Is an original parameter of HICUM, playing a fundamental role in the operation of the model, Q p，T Is the hole charge, V B’E’ Is the voltage between B - E, v B'C' Is the voltage between B - C, Vc is the collector voltage, Ib is the base current, V T Is the turn-on voltage. It can be seen from the formula that in the expression i Tf And i Tr There is only a voltage difference, but actually i Tf And i BE Are affected by the soft knee effect. When the actual effective base - emitter (B - E) voltage decreases at high currents, the ideality factor also has a certain deviation.

[0044] S2.2: Considering the influence of the Knee - Soft effect on the transfer current and the base - junction current, qks is a fitting model added for the Knee - Soft effect, that is, an empirical model for the decrease in current gain at high currents and low collector voltages. The formula of the added model becomes:

[0045]

[0046]

[0047] Among them, Kndc, Vc1, kb are newly added empirical parameters. The addition of these parameters is for better fitting of the model. The addition of the Vc1 fitting parameter is to fit the upward part of the latter half of the curve.

[0048] Furthermore, the empirical model qks of the INP HBT described in the formula is:

[0049]

[0050] Among them, ick is the critical current with a high bias effect, and Q p0 is the original parameter of HICUM, which plays a fundamental role in the operation of the model. Kn, Kv, Kn1, and Kndc1 are newly added empirical parameters. Kv is a correction coefficient related to the physical effect of voltage scaling, which is a coefficient used to modify the effective base voltage. The effect of the model is that it softens the bending behavior of the common-emitter I-V curve. The denominator of qks adds 1 to qk to avoid non-convergence of the calculation results due to the zero value of qks. A new charge qk is introduced into the collector current of this model to simulate the soft knee effect caused by the base widening effect observed in HBT during the high-level injection process of the collector, so as to more accurately reflect the actual influence of voltage on the physical quantities described by the entire formula. Among them, if the effect is not obvious, the added formula can be turned off, and Kndc is set to 0 and Kndc1 is set to 1.

[0051] S3: Based on the constructed INP HBT collector current DC model, obtain the initial values of the parameters related to the Knee-Soft effect, and simulate the INP HBT collector current DC model in the simulation software to obtain the simulation data of I C at different V B and I C .

[0052] Specifically, first, the initial values of the parameters unrelated to the Knee-Soft effect need to be obtained. The parameters unrelated to the Knee-Soft effect include: C10 and the zero-bias hole charge Q p,T that composes Q p0 , the B-E depletion charge weighting factor hjei in the heterojunction bipolar transistor, etc.; the parameters related to the Knee-Soft effect include: the fitting coefficients Kn, Kv, Kn1, and the parameter Kndc1 related to the degree of occurrence of the Knee-Soft effect; Kv is a correction coefficient related to the physical effect of voltage scaling, which can more accurately reflect the actual influence of voltage on the physical quantities described by the entire formula. Embed the collector current formula in step S2 into the INP HBT collector current DC model. With the initial values of the relevant and irrelevant parameters determined, obtain the simulation data of I C at different V B and I C . As shown in Figure 2 .

[0053] S4: Input the simulation data and the measured data into the large-signal model to establish the large-signal model, and fit and optimize the parameters in the INP HBT collector current DC model through the simulation data and the measured data to obtain the final INP HBT large-signal model.

[0054] S4.1: Establish the large-signal model.

[0055] Establish the large-signal model topology in IC-CAP, and embed the InP HBT collector current DC model into the large-signal model to form the final topology. Among them, the large-signal model topology is a non-linear model, so non-linear formulas are embedded. The modified InP HBT collector current DC model is carried out in two stages. That is, the formulas (1), (2), and (3) in the model are the formulas before modification. After considering their influence, the formulas in the model are modified into the forms of formulas (4), (5), (6), (7), and (8). Based on this, the first stage of the modified InP HBT collector current DC model is the HICUM model without the Knee-Soft effect, and the second stage is the HICUM model with the Knee-Soft effect. Fit and optimize the simulation results and measured results of the first stage in the simulation software IC-CAP to obtain the optimized parameters independent of the Knee-Soft effect.

[0056] S4.2: Fit the parameters.

[0057] During the parameter fitting process, the simulation software IC-CAP automatically generates the simulation data curve and the measured data curve. Adjust the initial values of the irrelevant parameters through the optimize control in the simulation software IC-CAP to make the simulation data curve infinitely close to the measured data curve, and record the irrelevant parameter values at this time, that is, the optimized parameters independent of the Knee-Soft effect. The adjustment of the parameters independent of the Knee-Soft effect is based on the cooperation among formulas (1), (2), and (3), and they are adjusted simultaneously to finally achieve the effect that the simulation data curve is close to the measured data curve.

[0058] Fit the simulation results and measured results of the second stage in the simulation software IC-CAP to obtain the optimized parameters related to the Knee-Soft effect, that is, the optimized relevant parameters.

[0059] Specifically, input the initial values of the relevant parameters obtained in step S4.1 into the simulation software IC-CAP. The simulation software IC-CAP automatically generates the simulation data curve and the measured data curve. Adjust the size of the initial values of the relevant parameters through the optimize control in the simulation software IC-CAP to make the simulation data curve approximately describe the change of the measured data curve, and record the relevant parameter values at this time, that is, the optimized relevant parameters. The adjustment of the relevant parameters is based on the cooperation among formulas (4), (5), (6), (7), and (8), and they are adjusted simultaneously to finally achieve the effect that the simulation data curve is close to the measured data curve.

[0060] In summary, the present invention constructs a large-signal model topology through the simulation software IC-CAP, embeds the collector current formula and the initial values of related parameters into the DC model of the InP HBT collector current, and simulates the DC model of the InP HBT collector current with the optimize control in the simulation software IC-CAP. The simulation data of the collector current-drain voltage under different gate voltages and the measured data are jointly simulated and fitted for optimization, so that the simulation data curve can accurately describe the change of the measured data curve, thereby determining the optimized related parameters and irrelevant parameters, and establishing the final large-signal model including the Knee-Soft effect, so as to accurately reflect the working state of the device.

[0061] The present invention also provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement a modeling method for a large-signal model of an InP HBT with a Knee-Soft effect. Those skilled in the art can understand that all or part of the processes for implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.

[0062] The present invention also provides a computer program, which is used to execute a modeling method for a large-signal model of an InP HBT with a Knee-Soft effect when the computer program is executed by a processor.

[0063] Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Claims

1. A modeling method of INP HBT large signal model with Knee-Soft effect, characterized in that The specific steps include: Obtain the measured data of INP HBT device for DC electrical characteristics test; Construct an INP HBT collector DC model, and obtain the initial values ​​of the parameters related to the Knee-Soft effect in the INP HBT collector DC model based on the measured data; simulate the INP HBT collector DC model to obtain the simulation data of the collector under different electrical signals; The INP HBT collector current DC model is embedded in the large signal model to establish the large signal model; the simulation data and the measured data are input into the large signal model, and the INP HBT collector current DC model performs parameter fitting optimization on the input data. If the fitting degree of all parameters meets the standard, the final INP HBT large signal model is obtained.

2. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 1, characterized in that: When constructing the INP HBT collector DC model, the influencing items include: current transmission influence and base junction current influence; The expression of the current transmission effect is: The expression of the base junction current influence is: Among them, i T is the total transmission current, i Tf is the forward transmission current, i Tr is the reverse transmission current; i jBEi is the internal base current, i jBEp is the external base current; i jBEi By the saturation current I BEiS and I REiS and the non-ideal coefficient m BEi and m REi Constructed; i jBEp By the saturation current I BEpS and I REpS and the non-ideal coefficient m BEp and m REp Composition; C 10 is the original parameter of the HICUM model, Q p,T is the hole charge, v B'E' is the voltage between BE, V T is the turn-on voltage, v B'C' is the voltage between BC, qks is the empirical model added for Knee-Soft effect, Kndc, Vc1, kb are newly added empirical parameters, Vc is the collector voltage, and Ib is the base current.

3. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 2, characterized in that: The expression of the empirical model qks is: Where qk is the new charge introduced in the collector current, which is used to simulate the soft knee effect caused by the base widening effect during the collector high-level injection process of the INP HBT collector current DC model; Q p0 are the original parameters of HICUM, Kn, Kn1, and Kndc1 are the newly added empirical parameters, and K v is the correction factor related to the physical effect of voltage scaling; C 10 is the original parameter of the HICUM model, V B’E’ is the voltage between BE, V T is the turn-on voltage, and ick is the critical current with high bias effect.

4. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 1, characterized in that: The initial values ​​obtained based on the influencing items of the INP HBT collector DC model also include: initial values ​​unrelated to the Knee-Soft effect; When constructing the INP HBT collector DC model, it is divided into a first stage and a second stage according to the Knee-Soft effect. The INP HBT collector DC model of the first stage includes the Knee-Soft effect, and the INP HBT collector DC model of the second stage does not include the Knee-Soft effect.

5. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 4, characterized in that: When the INP HBT collector DC model performs fitting optimization on the parameters of the first stage, parameters unrelated to the Knee-Soft effect are obtained; when the INP HBT collector DC model performs fitting optimization on the parameters of the second stage, parameters related to the Knee-Soft effect are obtained.

6. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 5, characterized in that: When the INP HBT collector DC model is fitted with parameters, First, select parameters unrelated to the Knee-Soft effect for fitting. If the fitting degree does not meet the standard, adjust the irrelevant parameters and fit again. If the fitting degree meets the standard, fit the parameters related to the Knee-Soft effect. If the fitting degree of some parameters related to the Knee-Soft effect does not meet the standard, the relevant parameters are adjusted and refitted. If the fitting degree of all relevant parameters does not meet the standard, the irrelevant parameters are adjusted and refitted. If the fitting degree of all relevant parameters meets the standard, the overall parameters are adjusted to obtain the INP HBT large signal model.

7. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 1, characterized in that: The electrical signal includes a collector voltage V C and base current I B .

8. The modeling method of the INP HBT large signal model with Knee-Soft effect according to claim 1, characterized in that: When establishing the large signal model, the topological structure of the large signal model is first constructed in the simulation software, and then parameter fitting is performed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the modeling method of the INP HBT large-signal model with Knee-Soft effect as claimed in any one of claims 1 to 8.

10. A computer program, characterized in that When the computer program is executed by a processor, it is used to execute the modeling method of the INP HBT large signal model with Knee-Soft effect as claimed in any one of claims 1 to 8.

Images