CN103364752B

CN103364752B - A kind of field calibration method in sheet load balance factor measuring system

Info

Publication number: CN103364752B
Application number: CN201310306006.7A
Authority: CN
Inventors: 栾鹏; 梁法国; 韩志国; 李静强; 孙晓颖; 孙静; 韩利华; 张贵军
Original assignee: CETC 13 Research Institute
Current assignee: CETC 13 Research Institute
Priority date: 2013-07-19
Filing date: 2013-07-19
Publication date: 2015-12-23
Anticipated expiration: 2033-07-19
Also published as: CN103364752A

Abstract

The invention discloses an on-site calibration method of an on-chip load pulling measurement system. In this calibration method, a mismatch attenuation monolith with a reflection coefficient covering 0.1 to 0.8 is first produced as a transfer standard, and then a vector network analyzer calibrated on the chip is used to calibrate and measure the transfer standard to obtain the standard conversion gain of the transfer standard G _T (S) and its calibration uncertainty, and then use the calibrated transfer standard parts to calibrate the conversion gain parameters of the on-chip load-pull measurement system, and complete the value traceability of the load-pull measurement system. This calibration technology can perform comprehensive and real calibration on the performance index of the on-chip load-pull system, give the calibration deviation △G _T , and can quantitatively give the measurement uncertainty of the on-chip load-pull measurement system, which is very important for the realization of load-pull The unification of the value of the measurement system is helpful, and it provides the technical support of measurement for the development and production of power monolithic circuits.

Description

A Field Calibration Method for On-Wire Load-Pull Measurement System

技术领域technical field

微波/毫米波测量领域，关于通过改变源或负载阻抗获得器件准确的大信号参数性能的技术。In the field of microwave/millimeter wave measurement, it is about the technique of obtaining accurate large-signal parameter performance of devices by changing the source or load impedance.

背景技术Background technique

对于线性器件而言，通过小信号下S参数可以推算出任何负载下的性能，负载牵引方法并不是必须的。但是微波功率晶体管输出功率大，一般工作在大信号状态下，表现出很强的非线性特性。因此，传统的基于线性理论的小信号设计方法已经无法满足大信号条件下微波功率晶体管的设计要求。各种功率放大器广泛应用于各种电子装备中，而功放的设计实际就是研制功率器件之间的匹配网络，研究什么样的网络能获得高的输出功率、高的效率和所需要的增益，这就要了解器件的输出功率、效益和增益随负载阻抗变化是如何变化的,负载牵引测量系统就能提供这些信息。设计者折中选择最佳匹配阻抗，设计出理想匹配网络，充分发挥器件的能力，获得高的输出功率、效率和所需增益。For linear devices, the performance under any load can be deduced through S-parameters under small signals, and the load-pull method is not necessary. However, microwave power transistors have a large output power and generally work in a large signal state, showing strong nonlinear characteristics. Therefore, the traditional small-signal design method based on linear theory can no longer meet the design requirements of microwave power transistors under large-signal conditions. Various power amplifiers are widely used in various electronic equipment, and the design of the power amplifier is actually to develop the matching network between power devices, and to study what kind of network can obtain high output power, high efficiency and required gain. To understand how the device's output power, efficiency, and gain vary with load impedance, a load-pull measurement system can provide this information. Designers compromise to choose the best matching impedance, design an ideal matching network, give full play to the ability of the device, and obtain high output power, efficiency and required gain.

负载牵引测量系统非常复杂，特别是在片测量系统，其整体性能指标并没有有效的计量措施，由于引进不同厂商的产品以及人员、条件等因素的制约，各单位提供的负载牵引参数测量结果差别很大，导致有时两套类似系统在同样的阻抗下测试输出功率、增益、效率的曲线也会不一致。为实现测量系统的量值准确可靠，保障负载牵引参数测量的一致性，为提高核心电子器件测试分析和设计开发能力提供有力的计量技术保障，因此，十分有必要开展在片负载牵引测量系统的校准工作。The load-pull measurement system is very complicated, especially in the sheet measurement system. There are no effective measurement measures for its overall performance indicators. Due to the introduction of products from different manufacturers and the constraints of personnel and conditions, the measurement results of load-pull parameters provided by each unit are different. It is very large, so sometimes the curves of two similar systems testing output power, gain, and efficiency under the same impedance will be inconsistent. In order to realize the accurate and reliable value of the measurement system, ensure the consistency of load pull parameter measurement, and provide a strong measurement technology guarantee for improving the core electronic device test analysis and design and development capabilities, it is very necessary to carry out the on-chip load pull measurement system. Calibration work.

发明内容Contents of the invention

本发明提供了一种在片负载牵引测量系统的现场校准方法，该方法是为了统一负载牵引测量系统参数量值，提供实用的负载牵引参数量值溯源途径所研制的。The invention provides an on-site calibration method of an on-chip load pulling measurement system, which is developed to unify the parameter values of the load pulling measurement system and provide a practical path for traceability of the load pulling parameter values.

为了解决上述技术问题，本发明采用的技术方案为：一种在片负载牵引测量系统的现场校准方法，包括如下步骤：In order to solve the above-mentioned technical problems, the technical solution adopted in the present invention is: an on-site calibration method of an on-chip load pulling measurement system, comprising the following steps:

第一步，研制一系列反射系数不同的失配衰减单片作为传递标准件，所述传递标准件的反射系数的覆盖范围为：0.1-0.8；The first step is to develop a series of mismatch attenuation monoliths with different reflection coefficients as transfer standard parts, and the coverage range of the reflection coefficient of the transfer standard parts is: 0.1-0.8;

第二步，利用矢量网络分析仪和微波探针台组成在片矢量网络分析仪作为定标装置，用定标装置测量各传递标准件的S参数，所述S参数为S₁₁、S₁₂、S₂₁、S₂₂；In the second step, use a vector network analyzer and a microwave probe station to form an on-chip vector network analyzer as a calibration device, and use the calibration device to measure the S parameters of each transfer standard. The S parameters are S ₁₁ , S ₁₂ , S ₂₁ , S ₂₂ ;

第三步，由设定的定标装置的源、负载端反射系数结合上一步测得的S参数，参照公式（1）计算得到各传递标准件的标准转换增益G_T（S），并将其作为相应传递标准件转换增益的定标标准值；In the third step, the standard conversion gain G _T (S) of each transfer standard is calculated by referring to the formula (1) based on the reflection coefficient of the source and load end of the calibration device and the S parameter measured in the previous step. It is used as the calibration standard value of the conversion gain of the corresponding transfer standard part;

${G G}_{T T} ((S S)) = = \frac{((11 - - {| | {Γ Γ}_{S S} | |}^{22})) {| | {S S}_{21 twenty one} | |}^{22} ((11 - - {| | {Γ Γ}_{L L} | |}^{22}))}{{| | ((11 - - {S S}_{1111} {Γ Γ}_{S S})) ((11 - - {S S}_{22 twenty two} {Γ Γ}_{L L})) - - {S S}_{1212} {S S}_{21 twenty one} {Γ Γ}_{S S} {Γ Γ}_{L L} | |}^{22}} . . . . . . ((11))$

其中，Γ_S为从传递标准件输入端向信号源端看去的反射系数；Γ_L为从传递标准件输出端向输出负载端看去的反射系数；Wherein, Γ _S is the reflection coefficient seen from the input end of the transfer standard part to the signal source end; Γ _L is the reflection coefficient seen from the output end of the transfer standard part to the output load end;

第四步，计算各传递标准件S参数基于定标装置的测量不确定度；The fourth step is to calculate the measurement uncertainty of the S parameters of each transfer standard part based on the calibration device;

第五步，通过偏微分方程得到由于S参数测不准引入的标准转换增益G_T（S）的测量不确定度，再将测量重复性引入的测量不确定度与之合成得到标准转换增益G_T（S）的定标不确定度；The fifth step is to obtain the measurement uncertainty of the standard conversion gain G _T (S) due to the uncertainty of the S parameter through the partial differential equation, and then combine the measurement uncertainty introduced by the measurement repeatability with it to obtain the standard conversion gain G Calibration uncertainty of _T (S);

第六步，用在片负载牵引测量系统测量各传递标准件的测量转换增益G_T，得到在片负载牵引测量系统测量重复性引入的测量不确定度，并将其与上述步骤得到的G_T（S）的定标不确定度合成得到在片负载牵引测量系统的测量不确定度，实现在片负载牵引测量系统的校准。The sixth step is to use the on-chip load-pull measurement system to measure the measurement conversion gain G _T of each transfer standard, to obtain the measurement uncertainty introduced by the measurement repeatability of the on-chip load-pull measurement system, and compare it with the G _T obtained in the above steps The calibration uncertainty of (S) is synthesized to obtain the measurement uncertainty of the on-chip load-pull measurement system to realize the calibration of the on-chip load-pull measurement system.

上述第一步中传递标准件的实现形式为：The implementation form of transferring standard parts in the first step above is:

设计上：采用稳定性好、分布参数小且适合版图布置的平衡式∏型电阻网络的形式；In terms of design: it adopts the form of balanced Π-type resistor network with good stability, small distribution parameters and suitable for layout layout;

工艺上：所述传递标准件以砷化镓晶圆片为衬底，利用溅射技术，把镍铬合金在衬底上制成高频电阻网络，高频电阻网络制成平衡式∏型电阻网络的形式，并在其上覆盖氮化硅薄膜；研制完成的电阻网络以微带线的形式在砷化镓圆片上制作，圆片经过减薄到100μm后作过孔处理；最后的传递标准件在圆片上保存。In terms of technology: the transfer standard part uses gallium arsenide wafer as the substrate, and uses sputtering technology to make a high-frequency resistor network on the nickel-chromium alloy on the substrate, and the high-frequency resistor network is made into a balanced Π-type resistor network, and cover it with a silicon nitride film; the developed resistance network is fabricated on a gallium arsenide wafer in the form of a microstrip line, and the wafer is thinned to 100 μm and then processed through holes; the final transfer standard The pieces are kept on the disc.

上述第四步的计算方法为：测量得到定标装置的系统剩余误差项及稳定性误差、噪声误差，并利用这些误差计算得到S参数的测量不确定度。The calculation method of the fourth step above is: measure the system residual error item, stability error, and noise error of the calibration device, and use these errors to calculate the measurement uncertainty of the S parameter.

上述第五步中标准转换增益G_T（S）的定标不确定度的计算方法为：The calculation method of the calibration uncertainty of the standard conversion gain G _T (S) in the fifth step above is:

首先，将各S参数分解为模和相位相结合的形式，即S₁₁＝r₁₁∠θ₁₁，S₂₁＝r₂₁∠θ₂₁，S₁₂＝r₁₂∠θ₁₂，S₂₂＝r₂₂∠θ₂₂，标准转换增益G_T（S）对各S参数的模和相位进行偏微分得到式（2）：Firstly, decompose each S parameter into the form of combination of mode and phase, that is, S ₁₁ =r ₁₁ ∠θ ₁₁ , S ₂₁ =r ₂₁ ∠θ ₂₁ , S ₁₂ =r ₁₂ ∠θ ₁₂ , S ₂₂ =r ₂₂ ∠θ 12 θ ₂₂ , the standard conversion gain G _T (S) performs partial differentiation on the modulus and phase of each S parameter to obtain formula (2):

$d d {G G}_{T T} ((S S)) = = \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{1111}} {dr dr}_{1111} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1111}} d d {θ θ}_{1111} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{21 twenty one}} {dr dr}_{21 twenty one} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{21 twenty one}} d d {θ θ}_{21 twenty one}$

$+ \frac{&PartialD; G_{T} (S)}{&PartialD; r_{12}} d r_{12} + \frac{&PartialD; G_{T} (S)}{&PartialD; θ_{12}} d θ_{12} + \frac{&PartialD; G_{T} (S)}{&PartialD; r_{22}} d r_{22} + \frac{&PartialD; G_{T} (S)}{&PartialD; θ_{22}} d θ_{22}$ ..................（9） $+ \frac{&PartialD; G_{T} (S)}{&PartialD; r_{12}} d r_{12} + \frac{&PartialD; G_{T} (S)}{&PartialD; θ_{12}} d θ_{12} + \frac{&PartialD; G_{T} (S)}{&PartialD; r_{twenty two}} d r_{twenty two} + \frac{&PartialD; G_{T} (S)}{&PartialD; θ_{twenty two}} d θ_{twenty two}$ ..................(9)

其中，dr₁₁、dr₂₁、dr₁₂、dr₂₂为四个S参数的模的不确定度，dθ₁₁、dθ₂₁、dθ₁₂、dθ₂₂为四个S参数的相位不确定度；Among them, dr ₁₁ , dr ₂₁ , dr ₁₂ , and dr ₂₂ are the uncertainties of the modes of the four S parameters, and dθ ₁₁ , dθ ₂₁ , dθ ₁₂ , and dθ ₂₂ are the phase uncertainties of the four S parameters;

接着，将由第四步得到的S参数的测量不确定度分解为模的测量不确定度和相位的测量不确定度；Then, decompose the measurement uncertainty of the S parameter obtained by the fourth step into the measurement uncertainty of the mode and the measurement uncertainty of the phase;

然后，根据模的测量不确定度和相位的测量不确定度，参照式（9）计算得到标准转换增益G_T（S）的测量不确定度；Then, according to the measurement uncertainty of the mode and the measurement uncertainty of the phase, the measurement uncertainty of the standard conversion gain G _T (S) is calculated by referring to formula (9);

最后，计算定标装置由于测量重复性引入的测量不确定度，将其与G_T（S）的测量不确定度合成得到标准转换增益G_T（S）的定标不确定度。Finally, calculate the measurement uncertainty introduced by the calibration device due to the measurement repeatability, and combine it with the measurement uncertainty of _GT (S) to obtain the calibration uncertainty of the standard conversion gain _GT (S).

本发明的设计思路如下：Design idea of the present invention is as follows:

图1所示为在片负载牵引测量系统的结构图。负载牵引测量系统具有设定源/负载端的反射系数的能力，通常负载牵引测量系统的测量不确定度主要来源于源/负载端的反射系数是否准确以及输出端功率测量是否准确，而这样的测量不确定度又不能直接测量出，因此就想到使用转换的方式间接测量得到不确定度。间接测量的物理量应该与源/负载端的反射系数相关，且该物理量中其他参数在负载牵引测量系统尽量不引入测量不确定度，这样此物理量所体现的测量不确定度才能比较准确的转换为负载牵引测量系统的测量不确定度。而转换增益G_T（S）正好符合这个要求。转换增益G_T（S）是与传递标准件本身的参数（四个S参数）及被校准测量系统的源阻抗Γ_s和负载阻抗Γ_L都有关的参量，S参数引入的负载牵引测量系统的测量不确定度很小，同时转换增益G_T还包含了对输出端功率测量结果的表征，因此选取该参数作为被校准系统的溯源参数能够合理评价负载牵引测量系统的测量性能，校准结果更加可信。Figure 1 shows the block diagram of the on-chip load-pull measurement system. The load-pull measurement system has the ability to set the reflection coefficient of the source/load terminal. Usually, the measurement uncertainty of the load-pull measurement system mainly comes from whether the reflection coefficient of the source/load terminal is accurate and whether the power measurement at the output terminal is accurate. The degree of certainty cannot be measured directly, so it is thought of using the conversion method to indirectly measure the degree of uncertainty. The physical quantity measured indirectly should be related to the reflection coefficient of the source/load end, and other parameters in the physical quantity should not introduce measurement uncertainty into the load pull measurement system as much as possible, so that the measurement uncertainty embodied in this physical quantity can be more accurately converted into load Measurement uncertainty of traction measurement systems. The conversion gain G _T (S) just meets this requirement. The conversion gain G _T (S) is a parameter related to the parameters of the transfer standard itself (four S parameters) and the source impedance Γ _s and load impedance Γ _L of the calibrated measurement system. The load-pull measurement system introduced by the S parameters The measurement uncertainty is very small, and the conversion gain G _T also includes the characterization of the output power measurement results, so choosing this parameter as the traceability parameter of the calibrated system can reasonably evaluate the measurement performance of the load pull measurement system, and the calibration results are more reliable. letter.

为了统一负载牵引测量系统参数量值，提供实用的负载牵引参数量值溯源途径，本发明的校准方法是通过设计制作性能稳定的不同反射系数（覆盖0.1到0.8）的失配衰减单片作为传递标准件，用在片校准过的矢量网络分析仪作为定标装置对传递标准件的标准转换增益G_T(S)进行定标，得到标准转换增益G_T(S)及其定标不确定度；再用在片负载牵引测量系统对定标后的传递标准件的转换增益G_T进行现场测量，将标准转换增益G_T(S)和测量转换增益G_T相比较得到△G_T，作为修正值，完成在片负载牵引测量系统量值溯源工作。图2为本发明的校准原理示意图。In order to unify the parameter value of the load-pull measurement system and provide a practical way to trace the source of the load-pull parameter value, the calibration method of the present invention is to design and manufacture mismatch attenuation single chips with different reflection coefficients (covering 0.1 to 0.8) with stable performance as the transmission For the standard part, use the on-chip calibrated vector network analyzer as the calibration device to calibrate the standard conversion gain G _T (S) of the transfer standard part, and obtain the standard conversion gain G _T (S) and its calibration uncertainty ; Then use the on-chip load-pull measurement system to measure the conversion gain G _T of the transfer standard after calibration, and compare the standard conversion gain G _T (S) with the measured conversion gain G _T to get △G _T as a correction value, and complete the traceability of the value of the on-chip load-pulling measurement system. Fig. 2 is a schematic diagram of the calibration principle of the present invention.

鉴于上面对校准方法的分析论述，采用上述技术方案取得的技术进步为：通过研制传递标准件并对其进行定标，可方便、准确的完成在片负载牵引测量系统的现场校准；本校准技术可以对在片负载牵引系统的性能指标进行全面、真实的校准，给出修正值△G_T，并能够定量的给出在片负载牵引测量系统的校准不确定度，实现工程应用中的不同在片负载牵引测量系统的测量结果量值准确、可靠，在片负载牵引测量系统经校准后进行器件设计的一致性得到显著提高；本发明中传递标准件的体积较小，采用标准的探针间距，适用于各种现有的测量系统，提高了本方法的普适性。In view of the above analysis and discussion of the calibration method, the technical progress achieved by adopting the above technical scheme is: through the development and transfer of standard parts and their calibration, the on-site calibration of the on-chip load pulling measurement system can be completed conveniently and accurately; this calibration The technology can comprehensively and truly calibrate the performance indicators of the on-chip load-pull system, give the correction value △G _T , and can quantitatively give the calibration uncertainty of the on-chip load-pull measurement system, realizing different engineering applications. The measurement results of the on-chip load-drawing measurement system are accurate and reliable, and the consistency of the device design is significantly improved after the on-chip load-drawing measurement system is calibrated; in the present invention, the volume of the transfer standard part is small, and a standard probe is used Spacing, applicable to various existing measurement systems, improves the general applicability of this method.

附图说明Description of drawings

图1为在片负载牵引测量系统的结构图；Figure 1 is a structural diagram of the on-chip load pulling measurement system;

图2为校准原理示意图；Figure 2 is a schematic diagram of the calibration principle;

图3为传递标准件设计结构示例图；Figure 3 is an example diagram of the design structure of transfer standard parts;

图4为定标装置的结构示意图。Fig. 4 is a schematic structural diagram of a calibration device.

具体实施方式Detailed ways

下面结合具体的实施方式对本发明进行更加详细的解释和说明。The present invention will be explained and described in more detail below in combination with specific embodiments.

一种在片负载牵引测量系统的现场校准方法，具体包括如下步骤：A method for on-site calibration of an on-chip load pulling measurement system, specifically comprising the following steps:

第一步，研制一系列不同反射系数的失配衰减单片作为传递标准件，所述传递标准件的反射系数覆盖范围为：0.1-0.8。In the first step, a series of mismatch attenuation monoliths with different reflection coefficients are developed as transfer standard parts, and the reflection coefficient coverage range of the transfer standard parts is: 0.1-0.8.

本发明是通过搭建定标装置对研制的传递标准件进行定标，统一在片负载牵引系统的量值，提供实用的在片负载牵引系统量值溯源途径，因此传递标准件的研制及定标工作是实现量值溯源的重要环节。The present invention calibrates the developed transmission standard parts by building a calibration device, unifies the value of the on-chip load traction system, and provides a practical way to trace the source of the value of the on-chip load traction system, so the development and calibration of the transmission standard parts Work is an important link to realize the traceability of quantity and value.

为保证在宽频带范围内具有稳定的、可重复的、受环境影响小的增益特性，提高本方法的准确度，需采用无源器作件为传递标准件。从传递标准件的溅射工艺及合金电阻的均匀性方面着手，设计制作在宽频带范围内符合性能要求的传递标准件。In order to ensure a stable, repeatable gain characteristic that is less affected by the environment in a wide frequency range and improve the accuracy of this method, passive devices must be used as transfer standard parts. Starting from the sputtering process of transfer standard parts and the uniformity of alloy resistance, design and manufacture transfer standard parts that meet performance requirements in a wide frequency range.

在设计上：以GaAs材料微带线的形式制作而成，采用平衡式∏型电阻网络的形式，可提高传递标准件的稳定性；为了减小分布参数并适合连接探针的间距要求，传递标准件的设计结构形式如图3所示：输入端和输出端的602欧姆电阻，可分别设计成两个1204欧姆的并联的形式减小分布参数。In terms of design: it is made in the form of GaAs material microstrip line, and adopts the form of balanced ∏-type resistance network, which can improve the stability of the transfer standard parts; in order to reduce the distribution parameters and meet the spacing requirements of the connecting probes, the transfer The design structure of the standard parts is shown in Figure 3: the 602 ohm resistors at the input and output ends can be designed as two parallel 1204 ohm resistors to reduce the distribution parameters.

在工艺上：标准件以砷化镓晶圆片为衬底，利用溅射技术，把镍铬合金在衬底上制成高频电阻网络，高频电阻网络制成平衡式∏型电阻网络的形式，如图3所示为衰减量为6dB、反射系数0.6的结构示意图，在其上覆盖氮化硅薄膜；研制完成的电阻网络以微带线的形式在砷化镓圆片上制作，圆片经过减薄到100μm后作过孔处理；最后的传递标准件就在圆片上保存。In terms of technology: the standard part uses gallium arsenide wafer as the substrate, and uses sputtering technology to make a high-frequency resistance network on the nickel-chromium alloy on the substrate, and the high-frequency resistance network is made into a balanced Π-type resistance network. Form, as shown in Figure 3 is a schematic diagram of the structure with an attenuation of 6dB and a reflection coefficient of 0.6, on which a silicon nitride film is covered; the developed resistor network is fabricated on a gallium arsenide wafer in the form of a microstrip line, and the wafer After being thinned to 100μm, the via hole is processed; the final transfer standard is saved on the wafer.

该发明制作的一系列传递标准件的技术性能如下：The technical performance of a series of transmission standard parts that this invention makes is as follows:

a、频率范围：2GHz-18GHz；a. Frequency range: 2GHz-18GHz;

b、反射系数分别为0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8；b. Reflection coefficients are 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8;

反射系数精度<±0.05Reflection coefficient accuracy <±0.05

c、衰减量为3dB时，反射系数分别为0.1,0.2,0.3；c. When the attenuation is 3dB, the reflection coefficients are 0.1, 0.2, 0.3 respectively;

衰减量为6dB时，反射系数分别为0.1,0.2,0.3,0.4,0.5；When the attenuation is 6dB, the reflection coefficients are 0.1, 0.2, 0.3, 0.4, 0.5 respectively;

衰减量为15dB时，反射系数分别0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8；When the attenuation is 15dB, the reflection coefficients are 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8;

衰减精度<0.1dBAttenuation accuracy <0.1dB

工艺参数：GSG间距为150μm，压点为50μm×50μm。Process parameters: the GSG spacing is 150 μm, and the pressure point is 50 μm×50 μm.

需要说明的是：传递标准件的定标精度取决于定标装置的准确度及传递标准件的稳定性，与传递标准件的设计指标（如反射系数<0.05）没有直接关系。It should be noted that the calibration accuracy of the transfer standard depends on the accuracy of the calibration device and the stability of the transfer standard, and has no direct relationship with the design index (such as reflection coefficient <0.05) of the transfer standard.

第二步，利用矢量网络分析仪和微波探针台组成在片矢量网络分析仪作为定标装置，如图4所示。用定标装置测量各传递标准件的S参数，所述S参数为S₁₁、S₁₂、S₂₁、S₂₂。In the second step, an on-chip vector network analyzer is used as a calibration device by using a vector network analyzer and a microwave probe station, as shown in Figure 4. Use a calibration device to measure the S parameters of each transfer standard, and the S parameters are S ₁₁ , S ₁₂ , S ₂₁ , and S ₂₂ .

在微波探针台上对矢量网络分析仪做TRL或LRM在片校准，以减小矢量网络分析仪的校准不确定度，这样在片矢量网络分析仪即可将端面校准到探针位置，实现了去嵌入校准，如图4所示。使用该定标装置测量各传递标准件的S参数，所述S参数包括S₁₁、S₁₂、S₂₁、S₂₂；Perform TRL or LRM on-chip calibration for the vector network analyzer on the microwave probe station to reduce the calibration uncertainty of the vector network analyzer, so that the on-chip vector network analyzer can calibrate the end face to the probe position, realizing de-embedding calibration, as shown in Figure 4. Using the calibration device to measure the S parameters of each transfer standard, the S parameters include S ₁₁ , S ₁₂ , S ₂₁ , S ₂₂ ;

其中，Γ_S为从传递标准件输入端向信号源端看去的反射系数；Γ_L为从传递标准件输出端向输出负载端看去的反射系数。Among them, Γ _S is the reflection coefficient viewed from the input end of the transfer standard to the signal source; Γ _L is the reflection coefficient viewed from the output end of the transfer standard to the output load end.

公式（1）经过变换可得到公式（2）：Formula (1) can be transformed into formula (2):

${G G}_{T T} ((S S)) = = \frac{((11 - - {| | {Γ Γ}_{S S} | |}^{22})) {| | {S S}_{21 twenty one} | |}^{22} ((11 - - {| | {Γ Γ}_{L L} | |}^{22}))}{{| | 11 - - {Γ Γ}_{in in} {Γ Γ}_{S S} | |}^{22} {| | 11 - - {S S}_{22 twenty two} {Γ Γ}_{L L} | |}^{22}} . . . . . . ((22))$

其中Γ_in为从传递标准件输入端向负载端看去的反射系数；Wherein Γ _in is the reflection coefficient seen from the input end of the transmission standard to the load end;

${Γ Γ}_{in in} = = {S S}_{1111} + + \frac{{S S}_{21 twenty one} {S S}_{1212} {Γ Γ}_{L L}}{11 - - {S S}_{22 twenty two} {Γ Γ}_{L L}} . . . . . . ((33))$

转换增益的定义为：传送到负载的功率与来自源的可资用功率之比，由此定义可知，当Γ_S与Γ_in共轭匹配时来自源的可资用功率是可以传送到网络的最大功率。Conversion gain is defined as the ratio of the power delivered to the load to the available power from the source. From this definition, it can be seen that when Γ _S and Γ _in conjugate match The available power from the source is the maximum power that can be delivered to the network.

在片负载牵引测量系统具有设定源/负载端的反射系数的能力，因此根据转换增益定义中传送到负载的功率，可任意设定Γ_L，对于无源传递标准件，当时由公式（2）可得到转换增益G_T（S）的变换形式公式（4），其实际就是器件的功率增益。The on-chip load-pull measurement system has the ability to set the reflection coefficient of the source/load end, so according to the power transmitted to the load in the conversion gain definition, Γ _L can be set arbitrarily. For the passive transfer standard, when At that time, formula (4) can be obtained from formula (2) to transform the conversion gain G _T (S), which is actually the power gain of the device.

${G G}_{T T} ((S S)) = = \frac{{| | {S S}_{21 twenty one} | |}^{22} ((11 - - {| | {Γ Γ}_{L L} | |}^{22}))}{((11 - - {| | {Γ Γ}_{in in} | |}^{22})) {| | 11 - - {S S}_{22 twenty two} {Γ Γ}_{L L} | |}^{22}} . . . . . . ((44))$

因此实际对在片负载牵引测量系统进行校准测量时，对于任意设定的负载阻抗Γ_L，都可以由公式（3）及快速算出相应的源阻抗Γ_s，而不需要通过牵引的办法找到对应的源阻抗Γ_s，并可以事先计算出传递标准件的相应标准转换增益G_T（S）。这对执行现场校准并给出校准结果的不确定度非常方便。Therefore, in the actual calibration measurement of the on-chip load-pull measurement system, for any set load impedance Γ _L , the formula (3) and Quickly calculate the corresponding source impedance Γ _s , without the need to find the corresponding source impedance Γ _s by means of traction, and calculate the corresponding standard conversion gain G _T (S) of the transfer standard in advance. This is very convenient for performing on-site calibrations and gives the uncertainty of the calibration results.

第四步，计算各传递标准件S参数基于定标装置的测量不确定度。The fourth step is to calculate the measurement uncertainty of the S parameters of each transfer standard based on the calibration device.

通过上面的分析可见：传递标准件的定标精度主要由定标装置的测量精度决定。为此，需测得定标装置的系统剩余误差项（反射跟踪误差、源匹配误差、方向性误差）及稳定性误差、噪声误差等，合成计算得到定标装置的测量不确定度。本方法将定标装置对S参数的测量不确定度作为传递标准件的测量不确定度的主要来源。下面以传递标准件在10GHz频率为例，计算定标装置对S参数的测量不确定度。It can be seen from the above analysis that the calibration accuracy of the transfer standard is mainly determined by the measurement accuracy of the calibration device. To this end, it is necessary to measure the residual error items of the calibration device (reflection tracking error, source matching error, directional error), stability error, noise error, etc., and synthesize and calculate the measurement uncertainty of the calibration device. In this method, the measurement uncertainty of the S parameter by the calibration device is used as the main source of the measurement uncertainty of the transfer standard part. Taking the transfer standard part at a frequency of 10GHz as an example, the measurement uncertainty of the calibration device for S parameters is calculated.

以S₁₁为例，不确定度的计算公式如下：Taking _S11 as an example, the calculation formula of uncertainty is as follows:

$Δ Δ {S S}_{1111 ((mag mag))} = = \sqrt{{Systematic Systematic}^{22} + + {Stability Stability}^{22} + + {Noise noise}^{22}} . . . . . . ((55))$

此处ΔS_11(mag)为S₁₁的测量不确定度，其中系统剩余误差Systematic²、稳定性及噪声引入的不确定度Stability²和Noise²均可根据Agilent公司PNA矢网误差产生公式求出。Here, ΔS _11(mag) is the measurement uncertainty of S ₁₁ , among which the system residual error Systematic ² , the uncertainty introduced by stability and noise Stability ² and Noise ² can be obtained according to the Agilent company PNA vector-network error generation formula .

系统剩余误差如式（6）所示：The system residual error is shown in formula (6):

${Systematic Systematic}^{22} = = {E E.}_{DF DF}^{22} + + {E E.}_{RF RF}^{22} {S S}_{1111}^{22} + + {E E.}_{XF XF}^{22} {S S}_{1111}^{44} + + {S S}_{21 twenty one}^{22} {S S}_{1212}^{22} (({E E.}_{LF LF}^{22} + + 44 {E E.}_{SF SF}^{22} {E E.}_{LF LF}^{22} {S S}_{1111}^{22} + + {E E.}_{LF LF}^{44} {S S}_{22 twenty two}^{22})) + + {A A}_{M m}^{22} {S S}_{1111}^{22} . . . . . . ((66))$

各项系统剩余误差可根据十二项误差的产生原理分别进行测量得到，结果见表1所示：The remaining errors of each system can be measured separately according to the principles of twelve errors, and the results are shown in Table 1:

表1片矢量网络分析仪系统剩余误差项测量结果Table 1 Measurement results of the residual error items of the vector network analyzer system

测试内容Test content 测量值（dB）Measured value (dB) 正向有效源匹配E_SF Positive active source match _ESF -32-32 反向有效源匹配E_SR Reverse Effective Source Matched E _SR -37-37 正向有效方向性E_DF Forward effective directivity E _DF -40-40 反向有效方向性E_DR Reverse effective directivity E _DR -43-43 正向有效负载匹配E_LF Positive payload matching E _LF -44-44 反向有效负载匹配E_LR Reverse payload matching E _LR -45-45 正向串扰E_XF Forward crosstalk E _XF -61-61 反向串扰E_XR Reverse Crosstalk E _XR -61-61 正向传输跟踪E_TF Forward Transport Tracking _ETF 0.060.06 反向传输跟踪E_TR Reverse Transmission Tracking E _TR 0.050.05 正向反射跟踪E_RF Forward Reflection Tracking E _RF 0.110.11 反向反射跟踪E_RR Retro-reflection tracking E _RR 0.040.04 幅度动态精度A_M Amplitude Dynamic Accuracy A _M 0.020.02 相位动态精度A_P（°）Phase dynamic accuracy A _P (°) 5.755.75

稳定性及噪声引入的不确定度也可通过测量并根据公式（7）、（8）求出：The uncertainty introduced by stability and noise can also be measured and calculated according to formulas (7) and (8):

${Stability Stability}^{22} = = \sqrt{{C C}^{22} + + {R R}^{22}} . . . . . . ((77))$

${Noise noise}^{22} = = {(({N N}_{T T} {S S}_{1111}))}^{22} + + {N N}_{F f}^{22} . . . . . . ((88))$

其中， $C^{2} = C_{RM 1}^{2} (1 + S_{11}^{4}) + {4 C}_{TM 1}^{2} S_{11}^{2} + C_{RM 2}^{2} S_{21}^{2} S_{12}^{2}$ in, $C^{2} = C_{RM 1}^{2} (1 + S_{11}^{4}) + {4 C}_{tm 1}^{2} S_{11}^{2} + C_{RM 2}^{2} S_{twenty one}^{2} S_{12}^{2}$

${R R}^{22} = = {(({R R}_{R R 11} ((11 + + {S S}_{1111}^{22})) + + {22 R R}_{T T 11} {S S}_{1111}))}^{22} + + {(({R R}_{R R 22} {S S}_{21 twenty one} {S S}_{1212}))}^{22}$

上式中除了S参数外，参数C_RM、C_TM为电缆稳定性分量；R_R1、R_T1为接头的连接重复性分量，N_T为迹线噪声；N_F为噪底。In the above formula, in addition to the S parameter, the parameters C _RM and C _TM are the cable stability components; R _R1 and R _T1 are the connection repeatability components of the connector, _NT is the trace noise; _NF is the noise floor.

S₂₁、S₁₂、S₁₂可根据Agilent公司PNA矢网误差产生公式按照上述步骤计算得出。S ₂₁ , S ₁₂ , and S ₁₂ can be calculated according to the above-mentioned steps according to the PNA vector-network error generation formula of Agilent Company.

第五步，通过偏微分方程得到由于S参数测不准引入的标准转换增益G_T（S）的测量不确定度，再将测量重复性引入的测量不确定度与标准转换增益G_T（S）的测量不确定度合成得到传递标准件标准转换增益G_T（S）的定标不确定度。The fifth step is to obtain the measurement uncertainty of the standard conversion gain G _T (S) due to the uncertainty of the S parameter through the partial differential equation, and then compare the measurement uncertainty introduced by the measurement repeatability with the standard conversion gain G _T (S ) is synthesized to obtain the calibration uncertainty of the standard conversion gain G _T (S) of the transfer standard.

利用定标装置测量S参数时不可避免的会引入不确定度，而S参数的不确定度会转嫁到标准转换增益G_T（S）上变成G_T（S）基于定标装置的测量不确定度。Uncertainty will inevitably be introduced when using a calibration device to measure S parameters, and the uncertainty of S parameters will be transferred to the standard conversion gain G _T (S) to become G _T (S) based on the measurement of the calibration device. Certainty.

标准转换增益G_T（S）的测量不确定度的计算方法为：The calculation method of the measurement uncertainty of the standard conversion gain G _T (S) is:

首先，将各S参数分解为模和相位相结合的形式，即S₁₁＝r₁₁∠θ₁₁，S₂₁＝r₂₁∠θ₂₁，S₁₂＝r₁₂∠θ₁₂，S₂₂＝r₂₂∠θ₂₂，标准转换增益G_T（S）对S参数的模和相位进行偏微分得到式（9）：Firstly, decompose each S parameter into the form of combination of mode and phase, that is, S ₁₁ =r ₁₁ ∠θ ₁₁ , S ₂₁ =r ₂₁ ∠θ ₂₁ , S ₁₂ =r ₁₂ ∠θ ₁₂ , S ₂₂ =r ₂₂ ∠θ 12 θ ₂₂ , the standard conversion gain G _T (S) performs partial differentiation on the modulus and phase of the S parameter to obtain formula (9):

$d d {G G}_{T T} ((S S)) = = \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; | | {S S}_{1111} | |} d d | | {S S}_{1111} | | + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1111}} d d {θ θ}_{1111} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; | | {S S}_{21 twenty one} | |} d d | | {S S}_{21 twenty one} | | + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{21 twenty one}} d d {θ θ}_{21 twenty one}$

$+ + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; | | {S S}_{1212} | |} d d | | {S S}_{1212} | | + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1212}} d d {θ θ}_{1212} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; | | {S S}_{22 twenty two} | |} d d | | {S S}_{22 twenty two} | | + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{22 twenty two}} d d {θ θ}_{22 twenty two}$

$= = \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{1111}} {dr dr}_{1111} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1111}} d d {θ θ}_{1111} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{21 twenty one}} d d {r r}_{21 twenty one} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{21 twenty one}} d d {θ θ}_{21 twenty one} . . . . . . ((99))$

$+ + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{1212}} d d {r r}_{1212} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1212}} d d {θ θ}_{1212} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{22 twenty two}} d d {r r}_{22 twenty two} + + \frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{22 twenty two}} d d {θ θ}_{22 twenty two}$

其中，dr₁₁、dr₂₁、dr₁₂、dr₂₂为四个S参数的模的不确定度，dθ₁₁、dθ₂₁、dθ₁₂、dθ₂₂为四个S参数的相位不确定度。Among them, dr ₁₁ , dr ₂₁ , dr ₁₂ , and dr ₂₂ are the uncertainties of the modes of the four S parameters, and dθ ₁₁ , dθ ₂₁ , dθ ₁₂ , and dθ ₂₂ are the phase uncertainties of the four S parameters.

接着，将上述第四步计算得到的S参数的测量不确定度分解为模的不确定度和相位的测量不确定度。Next, decompose the measurement uncertainty of the S parameter calculated in the fourth step above into the uncertainty of the mode and the measurement uncertainty of the phase.

然后，根据模的测量不确定度和相位的测量不确定度，参照式（9）计算得到标准转换增益G_T（S）的测量不确定度。Then, according to the measurement uncertainty of the mode and the measurement uncertainty of the phase, the measurement uncertainty of the standard conversion gain G _T (S) is calculated with reference to formula (9).

式（9）中，和的计算步骤如下：In formula (9), and The calculation steps are as follows:

首先，将式（9）转换为式（10）：First, convert formula (9) into formula (10):

${G G}_{T T} ((S S)) = = \frac{((11 - - {| | {Γ Γ}_{S S} | |}^{22})) {| | {S S}_{21 twenty one} | |}^{22} ((11 - - {| | {Γ Γ}_{L L} | |}^{22}))}{{| | 11 - - {S S}_{22 twenty two} {Γ Γ}_{L L} - - {S S}_{1212} {S S}_{21 twenty one} {Γ Γ}_{S S} {Γ Γ}_{L L} + + {S S}_{1111} (({S S}_{22 twenty two} {Γ Γ}_{S S} {Γ Γ}_{L L} - - {Γ Γ}_{S S})) | |}^{22}} . . . . . . ((1010))$

令Γ_S＝r_S∠θ_S，Γ_L＝r_L∠θ_L，σ＝(1-Γ_S|²)(1-|Γ_L|²)，R₁₁＝S₂₂Γ_SΓ_L-Γ_S＝r_R11∠θ_R11，P₁₁＝1-S₂₂Γ_L-S₁₂S₂₁Γ_SΓ_L，则式（10）可简化式（11）：Let Γ _S ＝r _S ∠θ _S , Γ _L ＝r _L ∠θ _L , σ＝(1-Γ _S | ² )(1-|Γ _L | ² ), R ₁₁ ＝S ₂₂ Γ _S Γ _L -Γ _S ＝r _R11 ∠θ _R11 ，P ₁₁ ＝1-S ₂₂ Γ _L -S ₁₂ S ₂₁ Γ _S Γ _L ， Then formula (10) can simplify formula (11):

${G G}_{T T} ((S S)) = = \frac{σ σ {r r}_{21 twenty one}^{22}}{{| | {P P}_{1111} + + {S S}_{1111} {R R}_{1111} | |}^{22}} = = \frac{σ σ {r r}_{21 twenty one}^{22}}{{| | \frac{{P P}_{1111}}{{R R}_{1111}} + + {S S}_{1111} | |}^{22} {| | {R R}_{1111} | |}^{22}} = = \frac{σ σ {r r}_{21 twenty one}^{22}}{{| | {Q Q}_{1111} + + {S S}_{1111} | |}^{22} {| | {R R}_{1111} | |}^{22}} . . . . . . ((1111))$

$= = \frac{σ σ {r r}_{21 twenty one}^{22}}{{r r}_{R R 1111}^{22} [[{r r}_{1111}^{22} + + {r r}_{Q Q 1111}^{22} - - {22 r r}_{1111} {r r}_{Q Q 1111} cos cos (({θ θ}_{1111} - - {θ θ}_{Q Q 1111}))]]}$

以的计算为例进行介绍，将式（11）对r₁₁和θ₁₁进行偏微分得到式（12）和式（13）：by The calculation of is introduced as an example, and formula (11) is partially differentiated with respect to r ₁₁ and θ ₁₁ to obtain formula (12) and formula (13):

$\frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {r r}_{1111}} = = - - \frac{{22 G G}_{T T}^{22} ((S S))}{σ σ {r r}_{21 twenty one}^{22}} {r r}_{R R 1111}^{22} [[{r r}_{1111} + + {r r}_{Q Q 1111} cos cos (({θ θ}_{1111} - - {θ θ}_{Q Q 1111}))]] . . . . . . ((1212))$

$\frac{&PartialD; &PartialD; {G G}_{T T} ((S S))}{&PartialD; &PartialD; {θ θ}_{1111}} = = \frac{{22 G G}_{T T}^{22} ((S S))}{{σr σr}_{21 twenty one}^{22}} {r r}_{R R 1111}^{22} {r r}_{1111} {r r}_{Q Q 1111} sin sin (({θ θ}_{1111} - - {θ θ}_{Q Q 1111})) . . . . . . ((1313))$

同理可推导出和将这些结果再代入到式（9）中即可得到G_T（S）的测量不确定度u₁。In the same way, it can be deduced and Substituting these results into formula (9) can get the measurement uncertainty u ₁ of _GT (S).

最后，由于定标装置测量不重复性还会引入测量不确定度u₂，因此计算得到u₂后将u₁与u₂合成得到标准转换增益G_T（S）的定标不确定度u_c1。u₂的计算方法是本领域技术人员都知晓的，这里不再赘述。Finally, since the measurement non-repeatability of the calibration device will also introduce the measurement uncertainty u ₂ , after calculating u ₂ , combine u ₁ and u ₂ to obtain the calibration uncertainty u _c1 of the standard conversion gain G _T (S) . The calculation method of u ₂ is well known to those skilled in the art, and will not be repeated here.

第六步，用在片负载牵引测量系统测量各传递标准件的转换增益G_T，得到在片负载牵引测量系统测量重复性引入的测量不确定度，并将其与上述步骤得到的G_T（S）的定标不确定度合成得到在片负载牵引测量系统的测量不确定度，实现在片负载牵引测量系统的校准。The sixth step is to use the on-chip load-pull measurement system to measure the conversion gain GT of each transfer standard to obtain the measurement uncertainty introduced by the repeatability of the on-chip load-pull measurement system, and compare it with the G _T obtained in the above steps ₍ The calibration uncertainty of S) is synthesized to obtain the measurement uncertainty of the on-chip load-pull measurement system, so as to realize the calibration of the on-chip load-pull measurement system.

在片负载牵引测量系统测量各传递标准件的转换增益G_T时，需要设定源/负载端的反射系数，这时，这些反射系数要与上述第三步中定标装置中设定的数值保持一致，这样才能保证校准的正确性。When measuring the conversion gain G _T of each transfer standard in the chip load pulling measurement system, it is necessary to set the reflection coefficient of the source/load end. At this time, these reflection coefficients should be kept with the values set in the calibration device in the third step above Consistent, so as to ensure the correctness of the calibration.

将在片负载牵引测量系统测量重复性引入的测量不确定度u₃与上述步骤得到的G_T（S）的定标不确定度u_c1合成得到在片负载牵引测量系统的测量测量不确定度u_c，实现在片负载牵引测量系统的校准。Combine the measurement uncertainty u ₃ introduced by the measurement repeatability of the on-chip load-pull measurement system with the calibration uncertainty u _c1 of G _T (S) obtained in the above steps to obtain the measurement measurement uncertainty of the on-chip load-pull measurement system u _c , to realize the calibration of the on-chip load-pull measurement system.

下面以一个具体的例子来详细说明在片负载牵引测量系统测量不确定度的计算方法。A specific example will be used to describe in detail the calculation method of the measurement uncertainty of the on-chip load-pulling measurement system.

以衰减量15dB、反射系数0.5的传递标准件在f=10GHz处的标准转换增益G_T（S）的定标不确定度为例进行在片负载牵引测量系统的测量不确定度评定。Taking the calibration uncertainty of the standard conversion gain G _T (S) at f=10GHz of the transfer standard with an attenuation of 15dB and a reflection coefficient of 0.5 as an example, the evaluation of the measurement uncertainty of the on-chip load-pull measurement system is carried out.

由定标装置测量S参数引入的不确定度即dG_T(S)：u₁＝0.186dB；The uncertainty introduced by the measurement of S parameters by the calibration device is dG _T (S): u ₁ =0.186dB;

由定标装置测量重复性引入的测量不确定度u₂＝0.050dB；The measurement uncertainty u ₂ ＝0.050dB introduced by the measurement repeatability of the calibration device;

两者合成得到该传递标准件的标准转换增益G_T（S）的定标不确定度为：Combining the two, the calibration uncertainty of the standard conversion gain G _T (S) of the transfer standard part is:

${u u}_{c c 11} = = \sqrt{{u u}_{11}^{22} + + {u u}_{22}^{22}} = = 0.193 0.193 dB dB$

用上述传递标准件在10GHz处，对0.8GHz-18GHz频段的在片负载牵引测量系统（Tuner型号：MT982BU）进行校准。利用该系统对转换增益G_T进行6次测量的结果如下（单位dB）：The on-chip load-pull measurement system (Tuner model: MT982BU) in the 0.8GHz-18GHz frequency band is calibrated at 10GHz with the above-mentioned transfer standard. The results of six measurements of conversion gain G _T using this system are as follows (in dB):

-15.486，-15.491，-15.495,-19.485，-15.481，-15.485，-15.486, -15.491, -15.495, -19.485, -15.481, -15.485,

由上述结果可以得到此系统由测量重复性引入的不确定度u₃＝0.035dB，From the above results, it can be obtained that the uncertainty u ₃ =0.035dB introduced by the measurement repeatability of this system,

则u_c1和u₃合成得到该系统的测量不确定度 Then u _c1 and u ₃ are combined to get the measurement uncertainty of the system

该系统的扩展不确定度U＝0.39dB（k=2）；The expanded uncertainty of the system U=0.39dB (k=2);

此即为在片负载牵引测量系统的测量不确定度最终结果。This is the final result of the measurement uncertainty of the on-wafer load-pull measurement system.

下述为本发明结果的验证及应用说明。The verification and application description of the results of the present invention are as follows.

利用本发明对0.8GHz-18GHz频段的在片负载牵引测量系统（Tuner型号：MT982BU）进行试校准。通过对反射系数分别为0.1、0.5、0.8的衰减量15dB的失配衰减器进行在片S参数测量，10GHz频点处校准结果见表2所示：The present invention is used for trial calibration of the on-chip load-pulling measurement system (Tuner model: MT982BU) in the 0.8GHz-18GHz frequency band. On-chip S-parameter measurements were performed on mismatch attenuators with reflection coefficients of 0.1, 0.5, and 0.8 with an attenuation of 15dB. The calibration results at the 10GHz frequency point are shown in Table 2:

表210GHz频点处在片负载牵引测量系统校准结果Table 2 Calibration results of the on-chip load-pull measurement system at the frequency point of 110 GHz

在片负载牵引测量系统并没有给出测量不确定度，通过本发明对系统的校准，就能够得到测量结果的不确定度。从标准转换增益G_T(s)和系统测量转换增益G_T的差值ΔG_T与不确定度的数值比较可以看出：该在片负载牵引测量系统测量结果量值比较可靠。在大反射系数（Γ＝0.8）时ΔG_T较大，分析原因主要有以下三点：The measurement uncertainty is not given in the sheet load pulling measurement system, and the uncertainty of the measurement result can be obtained by calibrating the system in the present invention. From the numerical comparison of the difference ΔG _T between the standard conversion gain G _T (s) and the system measured conversion gain G _T and the uncertainty, it can be seen that the measurement results of the on-chip load-pull measurement system are relatively reliable. When the reflection coefficient is large (Γ=0.8), ΔG _T is relatively large, and the reasons for the analysis mainly include the following three points:

1）在片负载牵引测量系统进行自校准时，大反射系数的S参数会产生较大不确定度；1) When the sheet load pulling measurement system is self-calibrating, the S parameter with a large reflection coefficient will have a large uncertainty;

2）传递标准件在大反射系数时进行S参数标定时也会产生较大不确定度；2) When the transfer standard is calibrated with a large reflection coefficient, the S parameter will also have a large uncertainty;

3）该负载牵引系统在片测量单片电路时，Γ＝0.8是其极限测量情况，系统不能够准确算出并找到最佳匹配点，会带来额外不确定度。3) When the load-pull system measures a single-chip circuit on a chip, Γ=0.8 is its limit measurement situation, and the system cannot accurately calculate and find the best matching point, which will bring additional uncertainty.

此外，还利用本发明的方法对另一台8GHz-50GHz在片负载牵引测量系统（Tuner型号：MT984AU）在10GHz处用Γ=0.5的传递件进行校准，以比较两套在片负载牵引测量系统的测量一致性。两系统的比较数据见表3所示：In addition, the method of the present invention is used to calibrate another 8GHz-50GHz on-chip load-pull measurement system (Tuner model: MT984AU) at 10GHz with a transfer element of Γ=0.5 to compare two sets of on-chip load-pull measurement systems measurement consistency. The comparative data of the two systems are shown in Table 3:

表3使用同一传递标准件对两套在片负载牵引测量系统测量数据比较Table 3 Comparison of measurement data of two sets of on-wafer load-pull measurement systems using the same transfer standard

参数parameter 0.8GHz-18GHz频段0.8GHz-18GHz frequency band 8GHz-50GHz频段8GHz-50GHz frequency band Γ_S(模/相角)Γ _S (mode/phase angle) 0.40∠10.260.40∠10.26 0.55∠10.970.55∠10.97 Γ_L(模/相角)Γ _L (mode/phase angle) 0.5∠-900.5∠-90 0.5∠-900.5∠-90 标准转换增益G_T(s)Standard Conversion Gain G _T (s) -15.48dB-15.48dB -15.46dB-15.46dB 系统测量转换增益G_T System measurement conversion gain G _T -15.54dB-15.54dB -15.74dB-15.74dB △G_T ΔG _T 0.06dB0.06dB 0.28dB0.28dB

从数据可知，该在片负载牵引测量系统在8GHz-50GHz频段的测量增益结果比0.8GHz-18GHz频段的测量结果偏小0.22dB。可见为了使各负载牵引测量系统量值准确可靠，可根据校准结果对测量系统进行相应修正来使各被校准系统具有更好的一致性，实现向上溯源的目的。It can be seen from the data that the measurement gain of the on-chip load-pull measurement system in the 8GHz-50GHz frequency band is 0.22dB smaller than the measurement result in the 0.8GHz-18GHz frequency band. It can be seen that in order to make the values of each load-pull measurement system accurate and reliable, the measurement system can be corrected according to the calibration results to make each calibrated system have better consistency and achieve the purpose of upward traceability.

通过验证示例及结果分析，本校准方法通过开展在片负载牵引测量系统的现场校准方法研究，并研制、标定传递标准件等途径解决了在片负载牵引测量系统的校准问题，对实现在片负载牵引测量系统的量值统一有所帮助，为功率单片电路的研制、生产提供了计量技术支撑。Through the verification example and result analysis, this calibration method solves the calibration problem of the on-wafer load-pull measurement system by carrying out the on-site calibration method research of the on-wafer load-pull measurement system, and develops, calibrates and transfers standard parts, etc. The unification of the value of the traction measurement system is helpful, and provides a technical support for the development and production of power monolithic circuits.

Claims

1. An on-site calibration method for an on-chip load traction measurement system, characterized by comprising the steps of:

firstly, developing a series of mismatched attenuation single chips with different reflection coefficients as a transmission standard component, wherein the amplitude coverage range of the reflection coefficients of the transmission standard component is as follows: 0.1-0.8;

secondly, an on-chip vector network analyzer composed of the vector network analyzer and the microwave probe station is used as a calibration device, multi-line TRL on-chip calibration is carried out on the vector network analyzer on the microwave probe station, and high-precision calibration of the calibration device is realizedAccurately, measuring S parameter of each transmission standard component by using a calibration device, wherein the S parameter is S₁₁、S₁₂、S₂₁、S₂₂；

Thirdly, calculating standard conversion gain G of each transmission standard component according to the set reflection coefficients of the source end and the load end and the S parameter measured in the last step and by referring to a formula (1)_T(S), and the standard value is used as the calibration standard value of the conversion gain of the corresponding transfer standard component;

is transformed

When in use_SAnd_inwhen the conjugation is matchedObtaining a conversion gain G_T(S)；

Wherein,_Sis the reflection coefficient as seen from the input end of the transfer standard towards the source end of the signal;_Lis the reflection coefficient seen from the output end of the transmission standard component to the output load end;

fourthly, calculating the measurement uncertainty of S parameters of each transmission standard component based on the calibration device;

fifthly, obtaining standard conversion gain G introduced by inaccurate measurement of S parameters through partial differential equation_T(S) and synthesizing the measurement uncertainty introduced by the measurement repeatability with the measurement uncertainty to obtain a standard conversion gain G_T(S) a scaling uncertainty; standard conversion gain G_TThe method for calculating the uncertainty of the measurement in (S) comprises the following steps:

first, each S parameter is decomposed into a form of a combination of a modulus and a phase, i.e., S₁₁＝r₁₁∠θ₁₁，S₂₁＝r₂₁∠θ₂₁，S₁₂＝r₁₂∠θ₁₂，S₂₂＝r₂₂∠θ₂₂Standard conversion gain G_T(S) partial differentiation is carried out on the mode and the phase of the S parameter to obtain an expression (9):

wherein dr is₁₁、dr₂₁、dr₁₂、dr₂₂Uncertainty of the mode, d θ, for four S parameters₁₁、dθ₂₁、dθ₁₂、dθ₂₂Phase uncertainty for four S parameters;

then, the measurement uncertainty of the S parameter obtained in the fourth step is decomposed into the measurement uncertainty of the mode and the measurement uncertainty of the phase;

then, according to the measurement uncertainty of the modulus and the measurement uncertainty of the phase, reference formula (9) adopts ingenious conversion element transformation to eliminate the influence of complex quantity, and the calculation of the modulus and the phase partial differential of the conversion gain is feasible, so that the measurement uncertainty of the standard conversion gain GT (S) is obtained;

in the formula (9), the reaction mixture is, andthe calculation steps are as follows:

first, formula (9) is converted to formula (10)

Order to_S＝r_S∠θ_S，_L＝r_L∠θ_L，σ＝(1-|_S|²)(1-|_L|²)，R₁₁＝S₂₂ _S _L-_S＝r_R11∠θ_R11，P₁₁＝1-S₂₂ _L-S₁₂S₂₁ _S _L，Then (10) can simplify formula (11):

to be provided withBy way of example, the calculation of (11) for r₁₁And theta₁₁Partial differentiation is carried out to obtain formula (12) and formula (13):

the same can be deducedAndthese results are substituted into the formula (9) to obtain G_T(S) measurement uncertainty u₁；

Finally, the measurement uncertainty introduced by the calibration device due to measurement repeatability is calculated,it is reacted with G_T(S) synthesizing the measurement uncertainty to obtain a standard conversion gain G_T(S) a scaling uncertainty;

sixthly, measuring the measurement conversion gain G of each transmission standard component by using the on-chip load traction measurement system_TObtaining the measurement uncertainty introduced by the measurement repeatability of the sheet load traction measurement system, and comparing the measurement uncertainty with the G obtained in the step_TAnd (S) synthesizing the calibration uncertainty to obtain the measurement uncertainty of the on-chip load traction measurement system, thereby realizing the calibration of the on-chip load traction measurement system.

2. The method of claim 1, wherein the standard transfer element is implemented in the form of:

in the design: a balanced n-shaped resistor network form which is good in stability, small in distribution parameter and suitable for layout arrangement is adopted;

the process is as follows: the standard transmission piece takes a gallium arsenide wafer as a substrate, utilizes a sputtering technology to make a nickel-chromium alloy into a high-frequency resistor on the substrate, and covers a silicon nitride film on the high-frequency resistor; the developed circuit is manufactured on a gallium arsenide wafer in a microstrip line mode, and the wafer is subjected to through hole processing after being thinned to 100 microns; the final transfer standard is stored on the wafer.

3. The on-site calibration method for the on-chip load traction measurement system according to claim 1, wherein the calculation method of the fourth step is as follows: and measuring to obtain a system residual error item, a stability error and a noise error of the calibration device, and calculating to obtain the measurement uncertainty of the S parameter by using the errors.