WO2008066137A1

WO2008066137A1 - Electronic part high-frequency characteristic error correction method and device

Info

Publication number: WO2008066137A1
Application number: PCT/JP2007/073110
Authority: WO
Inventors: Naoki Fujii; Gaku Kamitani; Taichi Mori
Original assignee: Murata Manufacturing Co., Ltd.
Priority date: 2006-11-30
Filing date: 2007-11-29
Publication date: 2008-06-05

Abstract

Provided is an electronic part high-frequency characteristic error correction method capable of performing calibration of a 2-terminal impedance part while a measurement system to be corrected is in the same state as in the actual measurement. At least three correction data acquiring samples having different high-frequency characteristics are measured in a reference measurement system and in an actual measurement system. An expression for correlating a measurement value obtained in the actual measurement system with a measurement value obtained in the reference measurement system is decided by using a transmission path error correction coefficient. An arbitrary electronic part (2) is measured in the actual measurement system and an estimated value of the high-frequency characteristic of the electronic part which would be obtained by measuring the electronic part in the reference measurement system is calculated by using the decided expression.

Description

Specification

Correction method for high frequency characteristics error of electronic parts

Technical field

[0001] The present invention relates to a method for correcting high-frequency characteristic errors of electronic components, and more particularly to a method for correcting errors in a measurement system in measuring high-frequency characteristics of a two-terminal impedance component. Background art

Conventionally, in the mass production process of electronic parts, the electrical characteristics of the electronic parts have been measured using an automatic characteristic sorter. Since the measurement system with the automatic characteristic sorter has different circuit characteristics from the standard measurement system, by correcting the measurement value with the automatic characteristic sorter and estimating the measurement value with the standard measurement system, Yield can be improved. As a method for performing such correction, techniques called SOLT, TRL calibration, and RRRR / TRRR calibration are known.

[0003] First, TRL / SOLT calibration will be explained!

[0004] TRL calibration is the most effective technique that can be used to measure the true value of the scattering coefficient matrix of a surface-mounted component that is the subject. Another widely used conventional technology is SOLT calibration. These will be briefly described.

[0005] In order to obtain the true value of the subject, it is necessary to identify the error factor of the measurement system and remove the influence of the error factor from the measurement result. Figure 1 shows the error model used in a typical error elimination method (calibration method).

That is, as shown in FIG. 1 (a), the electronic component 2 as the subject is connected on a transmission path formed on the upper surface of the measurement jig 10. Connectors 51a and 61 provided at one end of the coaxial cables 50 and 60 are connected to the connectors 11a and ib provided at both ends of the transmission line of the measuring jig 10, and the other ends of the coaxial cables 50 and 60 are Connected to the network analyzer measurement port (not shown). Arrows 51s and 61s indicate the calibration plane.

[0007] Figure 1 (b) is an error model for TRL correction, expressed as scattering coefficients S 1, S 2, S 3, and S 2.

11A 12A 21A 22A The scattering coefficient e, e, e, between the measurement jig circuit 12 and the terminal pairs a — b, a — b

1 1 2 2 00 01 10 e expressed as one measurement port side circuit 52 and scattering coefficient expressed as f, f, f, f The circuit 62 on the other measurement port side is connected.

[0008] Figure 1 (c) shows the error of the SOLT correction, which is measured by the scattering coefficients S 1, S 2, S 3, and S 2.

11A 12A 21A 22A

One measurement port represented by scattering coefficients E, E, Ι, Ε on both sides of the fixture circuit 14

DF RF SF

Circuit 54 on the other side and circuit 64 on the other measurement port side represented by the scattering coefficients Ε and Ε

LF TF

It has been continued.

[0009] In the case of SOLT calibration, in order to identify the cause of error, measurement must be performed by attaching a device with at least three types of scattering coefficients known to the subject measurement surface, as shown in Figure 2. Traditionally, open standard (ΟΡΕΝ), short circuit (SHORT), and termination (LOAD = 50 Q) standard devices 80, 81, 82 are often used. It is extremely difficult to realize the “terminal” standard device, and calibration cannot be performed at the tip of the jig 10 (the calibration surface indicated by arrows 51s and 61s). In a coaxial environment, such a standard can be realized by a method such as sliding road, so this method is widely used and is called SOLT calibration.

[0010] In TRL calibration, instead of a standard device with a difficult device shape, several types of transmission lines (Lines) with different lengths are used, with direct connection between ports (Through), total reflection (Reflection usually short-circuited), and different lengths. Used as a standard device. The transmission path of the standard device can be expected to be the most accurate, especially in a waveguide environment, as soon as a relatively known scattering coefficient is manufactured and the total reflection is short-circuited. Known as a high calibration method

[0011] Figure 3 shows the TRL calibration error factor derivation method. In the figure, the transmission line is hatched. The calibration surface is the connection with the device as shown by arrows 2s and 2t. In order to identify the error factors, the board 86 directly connected between the ports (Through), the board 83 of total reflection (normally short-circuited), and boards 84 and 85 of several types of transmission lines (Lines) of different lengths are used. Use as a standard device. In this example, Through is so-called Zero-Through. When measuring the subject, the subject 2 is connected in series to a measurement substrate 87 that is longer by the size of the subject.

[0012] The outline of TRL and SOLT calibration is as described above. Force S, These technologies have the following two problems.

[0013] (1) Coaxial transmission line, etc. (several types of lines and reflection) and Through -Calibration error will occur if the error factors that occur in the plane transmission line connection are not all equal. Even if the same type of connector is used for each standard device, the effect of connector characteristic variation becomes very large, especially if they are different, and it is practically impossible to approach the millimeter wave band.

[2] (2) In order to solve the above-mentioned problems, commercially available jigs have been devised to avoid the influence of variations in connector measurement by using a coaxial connector as a common contact connection with a standard device transmission path. However, since the coaxial pin is damaged, it is often difficult to secure a sufficient pressing load on the contact area. In addition, when the measurement frequency is increased, the transmission path and the coaxial pin are generally narrower, and the measurement variation due to these positioning repeatability increases.

[0015] In order to solve these problems, a so-called RRRR / TRRR calibration method has been proposed.

[0016] Next, an outline of the RRRR / TRRR calibration method will be described.

[0017] By short-circuiting the signal conductor and the ground conductor at predetermined power points on only one transmission path, these identify errors in the measurement system up to the front end of the transmission path. The characteristic is that the gas characteristics can be measured with high accuracy. The problem with TRL / SOLT calibration methods is that they are unrelated to variations in transmission path characteristics and variations in the contact state between the transmission path and coaxial connector.

The error model is the same as that of SOLT / TRL calibration as shown in FIGS. In other words, Fig. 4 shows the TRRR calibration error model, which is the same as the SOLT calibration error model shown in Fig. 1 (c). Figure 5 shows the RRRR calibration error model, which is the same as the TRL calibration error model shown in Figure 1 (b).

[0019] The point of the RRRR / TRRR calibration method is the measurement method of “standard measurement value” used for calibration. The measurement value of the standard device in SOLT and the standard transmission line in TRL is “standard measurement value”. In the RRRR / TRRR calibration method, as shown in Fig. 6, the measurement value measured by changing the position of the short-circuit reference on the measurement substrate 10a is taken as the “standard measurement value”. Since there is no influence of the connector, it can be said that this is a more accurate and effective method for desktop measurement than SOLT calibration and TRL calibration.

[0020] However, in TRRR / RRRR calibration, the jig transmission line 10s, 10t is connected to the short circuit reference (short chip). (2s) is used as a calibration reference because of the change in the reflection coefficient caused by the difference in the connection position, so if the wavelength of the signal to be measured is long (the frequency is low), the connection position of the short-circuit reference must be changed significantly. Since T and T in the figure need to be lengthened,

1 2

It is necessary to increase the length of the plate 10a (the dimension in the direction indicated by the arrow L). In addition, the automatic characteristic sorter used in the mass production process has restrictions on the structure and dimensions. Therefore, the jig 10a should be provided with a GND terminal for correction, and the short chip 2s can be positioned accurately. (For example, see Patent Documents 1 and 2).

Patent Document 1: WO2005 / 101033 Publication

Patent document 2: WO2005 / 101034

Non-Patent Document 1: Application Note 1287-9: In- Fixture Measurements Using Vector Net work Analyzers, ((1999 1999 Hewlett-Packard Company)

Disclosure of the invention

Problems to be solved by the invention

[0021] In the automatic characteristic sorter used in the mass production process of electronic parts, for example, as shown in the main part configuration diagram of Fig. 9, the test pins 32a and 32b protruding from the measurement terminal part 30 are connected to the test object. The electrodes 2a and 2b of an electronic component 2 are pressed against each other and connected in series between the measurement pins 32a and 32b, and the measurement pins 32a and 32b are connected to the measuring device (not shown) via coaxial cables 34 and 36. It is connected to the. If the space is small enough to connect the electronic component 2 around the measurement terminal section 30 and it cannot be secured, the sample on the measurement terminal section 30 is substantially the same as the mass production device itself or the mass production device. Measurement system error correction must be performed under the restriction that it cannot be connected. In such a case, the following problem arises.

[0022] (1) It is impossible to connect transmission lines of different lengths to the measurement terminal section of the automatic characteristic sorter, and TRL calibration cannot be applied.

[0023] (2) SOLT calibration cannot be calibrated at the end of the measurement terminal in practice, and is limited to being applied only to a coaxial or waveguide system. Normally, the coaxial connector is calibrated by SOLT calibration, and the subsequent transmission line is designed so that no error occurs, so that sufficient measurement accuracy is obtained. However, in the measurement terminal section of the force automatic characteristic sorter, there are restrictions on the shape and dimensions of the transmission line after the coaxial connector, so calibration alone to the coaxial connector section is sufficient. There are many cases where accuracy cannot be obtained.

[0024] (3) The following problems arise when trying to measure the standard device at the tip of the measurement terminal by using some means in SOLT calibration.

[0025] i) SOLT calibration requires measurement of 1-port devices at each port. That is, as shown in the plan view of the measurement board 10b in Fig. 7, when measuring two-terminal electronic parts in series connection between the slits 10k of one signal line ΙΟχ, it is not necessary for the measurement and is grounded to the terminal part. The terminal is V. However, since one-port devices cannot be measured without a ground conductor in SOLT calibration, it is necessary to provide a ground terminal only for calibration in order to apply SOLT calibration.

[0026] ii) The SOLT calibration requires measurement of three types of 1-port devices with known values at each of the two ports. However, as shown in the plan view of the measurement board 10c in FIG. Since the two terminals of the device are connected to the ground conductor 10g, it is impossible to measure a single-port device in which each port is independent.

[4] (2) When measuring 2-terminal electronic components in series connection, there is no ground terminal in the terminal section because it is not necessary for measurement. Since RRRR calibration requires short-circuit measurement, to apply RR RR calibration, it is necessary to provide a ground terminal only for calibration.

[0028] (5) In RRRR calibration, short circuit is performed at several points on the board! /, But when the frequency is low V, it is necessary to change the connection position of the short circuit reference significantly, and therefore the length of the measurement board It is necessary to lengthen the length.

[0029] In view of such circumstances, the present invention corrects the high-frequency characteristic error of an electronic component that can be calibrated for a two-terminal impedance component while the measurement system to be corrected remains in the same state as at the time of actual measurement. Is to provide a method.

Means for solving the problem

In order to solve the above-described problems, the present invention provides a method for correcting a high-frequency characteristic error of an electronic component configured as follows.

[0031] The method of correcting the high-frequency characteristic error of an electronic component may be obtained by measuring the electronic component, which is a two-terminal impedance component, using an actual measurement system and measuring the electronic component using a reference measurement system. This is a method of calculating an estimated value of the high frequency characteristics of the electronic component. Electric The sub-component high-frequency characteristic error correction method includes: (1) a first step of preparing at least three first correction data acquisition samples with different high-frequency characteristics, which are priced in the reference measurement system; ) At least three first correction data acquisition samples, or at least three second correction data acquisition samples that can be regarded as having the same high frequency characteristics as the first correction data acquisition sample, And (3) pricing data in the reference measurement system of the first correction data acquisition sample prepared in the first step and the actual measurement in the second step. From the measurement data of the first correction data acquisition sample or the second correction data acquisition sample measured in the system, the measurement value measured in the actual measurement system and the measurement value measured in the reference measurement system The A third step of determining an equation to be associated using an error correction coefficient of the transmission path; (4) a fourth step of measuring an arbitrary electronic component by the actual measurement system; and (5) the fourth step. Based on the measurement result obtained in step 3, the above formula determined in the third step is used to measure the electronic component in the reference measurement system. And a fifth step of calculating an estimated value of the high frequency characteristics of the electronic component.

[0032] The first correction data acquisition sample prepared in the first step may be pre-valued by other methods even if it is pre-valued by actually measuring with a reference measurement system. May be. For example, for a large number of samples that can be regarded as equivalent characteristics, only some of the samples may actually be measured with a reference measurement system, and the measured values may be used for pricing other samples.

[0033] According to the above method, the first and second steps can be executed using the correction data acquisition sample having substantially the same shape and size as the electronic component. Previously, the force that could only be calibrated up to the tip of the coaxial connector in the measurement system of the automatic characteristic sorter The compensation up to the tip of the terminal to which the electronic component is connected can be performed by the above method.

[0034] In a preferred aspect, in the actual measurement measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample and the second correction data acquisition are used. The sample for use and the electronic component are connected in series. The above formula is expressed in terms of terminals 1 and 2 where impedance Z is measured when an electronic component is measured by the reference measurement system, and

Terminals 1 and 2 where impedance Z is measured when electronic components are measured using the actual measurement system described in mmm Are derived on the basis of an error model connected between and. When calculating the impedance seen from the terminal 1 force, the error model is such that impedances Z and Z are connected in series between the terminal 1 and the terminal 1, and the impedance Z and Z An impedance Z is connected between the connection point and the ground, an impedance Z is connected between the terminal 2 and the terminal 2, and an impedance Z is connected between the terminal 2 and the ground. The impedances Z, Z, Z, Z, Z are

As a result of measuring the impedance of the first correction data acquisition sample, Z, Z, Z, and in the second step, at least three of the first correction data acquisition sample or the second correction data For the sample for acquisition, using the results Z, Z, Z of the impedance of the terminal 1 and the results of measuring the impedance of the terminal 2, ZZ, Z,

la]

denom ii (Z _d2 Z _d ) Z _m + (Z _{J {} ― Z _d3 ) Z _ml2 + (Z _d3

^ ll― [Sat ^ Z _d2 ― Ζ _ά jZ _d3i ― Z _d] ^ jZ _d3 one ^ jZ _mU one

— Dl) ^ ml ^ ⁺ (^ dl Z _m "Z _ml3 + (¾ ₂ ― Z _dl ) Z _mU Z _mU }]

I denom

= Sat

/ denom The following formula [number lb]

[Several lb]

de _om- (Z _d2 -Ζ ^) Ζ _Μ23 + (Z _d -Z _d3 ) Z _m22 + (Ζ _{ί / 3} -Z _dl ) Z _m2X = (土一 di Λ one ^ άΐ) Z _d3 ― Z _d2 ^ jZ _m22 ― Z _m2l ^ jZ _m23 ― ^ Z _m23 I I denom Z _d2 ~

X mu + one

Of the 16 Z, Z, Z, Z thread combinations obtained from I denom,

For the following mathematical model 2], it is determined by using at least one combination of Z, Z, and Z. [Equation 2]

Z _fl =-[{(Z ₂₂ + (Z ₂₁ + Z ₀ )) Z _dl + ((Z ₂₁ + Z.) + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _m

+ {(-Z ₁₂ -Z _U ) Z ₂₂ + (-Z, ₂ -Z _U ) (Z ₂ , + Z ₀ )} Z _dl

+ {(-Z ₁₂ -Z _H ) (Z ₂₁ + z.) — Z „Z _{12 22} -Z _n Z ₁₂ (Z ₂₁ + Z ₀ )]

/ [{(+ (Z ₂₁ + Z ₀ )} Z _mlI + (-Z _l2 -Z „) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z ₂₁ + Z ₀ )]

Z _{/ 2} =-[{(Z ₂₂ + (Z ₂₁ + Z.)) Z _d2 + ((Z ₂₁ + Z ₀ ) + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _ml2

+ {(-z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z,,) (Z ₂₁ + Z.)} Z _dl

+ ((— Zj ₂ -Λ _Π Ζ ₂₁ + z, ₀ ) -Z _{u 12} } Z ₂₂ -Z _n Z ₁₂ ^ ₂₁ + Z ₀ ) J

_{/ [{(Z 22 + (} Z 21 + Z 0)} Z ml2 + (- Z 12 -Z n) Z 22 + (- z 12 -ζ η χζ 2Ί + z).]

Z ₃ =-[{(Z ₂₂ + (Z ₂₁ + Z.)) Z _rf3 + ((Z ₂₁ + Z ₀ ) + Z ₁₂ ) Z ₂₂ + Z _n (Z ₂₁ + z ₀ )} z _ml3

+ {(-Z ₁₂ -Z _U ) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z _2l + Z ₀ )} Z _d3

+ {(-z ₁₂ -Z _n ) (Z ₂₁ + Z.)- _{2 22} -Z _n Z ₁₂ (Z ₂₁ + Z ₀ )]

/ [{(Z ₂₂ + (Z ₂₁ + Z ₀ )} Z _ml3 + (-Z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z _U ) (Z ₂₁ + Z ₀ )]

[0035] In this case, the measurement result in the reference measurement system can be estimated by correcting the error in the transmission path for the measurement result in the series connection actual measurement system.

[0036] In another preferred embodiment, in the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample and the second correction data acquisition The sample for use and the electronic component are shunt connected. The above formulas are terminals 1 and 2 at which admittance Y is measured when measuring an electronic component in the reference measurement system, and terminals at which admittance Y is measured when measuring an electronic component in the above-mentioned mmm actual measurement system. Derived based on an error model connected between 1, 2 and ddd. Admittan m seen from the terminal 1 force m

When the error model is derived, an admittance m d between the terminal 1 and the terminal 1

Y is connected, and a signal is connected between the connection point between the terminal 1 and the admittance Υ and the ground.

12 m 12

Admittance Y is connected, and the connection point between the admittance Y and the terminal 1 is connected to the ground.

11 12 d

An admittance Y is connected between the terminals 2 and 2, and an admittance Y is connected between the terminals 2 and 2, and an admittance Y is connected between the connection point of the admittance Y and the terminal 2 and the ground.

22 m

Y is connected. The admittance Y, Y, Y, Y, Y is the first step

21 f 11 12 21 22

Y, Υ, Y as a result of measuring the admittance of at least three of the first correction data acquisition samples in step 2, and in the second step, at least three of the first correction data As a result of measuring the admittance of the terminal 1 for the acquisition sample or the second correction data acquisition sample, 補正, 用い, Y, and the result of measuring the admittance of the terminal 2, Y, ,, Υ are used. And the following formula 3a]

[Number 3a]

_{denom = {Y d2 -) +} (-) Y ml2 + (- Y d2)

Tjl = ± [Satichi] ― One ^ Λ Μ

~ 1

^ Xdl

I denom

Y _X2 = 士 (】 ¾ ^{+ +} ) (— ll X li)

I denom The following formula [Equation 3b]

[ _Equation 3b] denom = (Y _d2 -Y _dl ) Y _m23 + (1) Y _m22 + (1 Y _d2 ) Y _m2i Sat] rfl Λ / 1]

-{(+ (-l)

I denom

Y ₂₂ = Sat ((Xd + (1)) (— X —

Of the 16 Y, ，, Υ, and Υ thread combinations obtained from I denom,

For the following formula [Equation 4]

[Equation 4]

Y _fl = ₂ + (7 ₂₁ + Γ ₀ )) ¾10 ((Y ₂₁ + Υο) + Yn ₂ 2 + Άι + u

_{_{+ {(- 12 -γ η)}} γ 22 + (-Y -Y U) (Y 21 + Y 0)} Y DL

+ -Y _n ) (Y _2] + r ₀ ) -7 _n 7 ₁₂ } 7 ₂₂ -F _li r _i2 (7 _{21 +} 7 ₀ )]

/ [{(¾ + (F ₂₁ + Y ₀ )} Y _mn + (-Y _u -Y Y22 + (-¾-+ r ₀ )] r _f2 =-[{(+ (¾ + Yo W _d2 + ( (+ Y ₀ ) + Yn) ¾ + Yn + ^)} ₂

Tens {(— ¾ -Y) Y ₂₂ + (-Y _l2

+ {(--) + ₀ ) -7 „7 ₁₂ } 7 ₂₂ -Y _n Y _l2 (Y + Y ₀ )]

/ [{(¾ + (Y _2l + Y ₀ )} Y _m n + (--+ (-

Y _f , =-+ (Y _2l + Y ₀ )) + ((+) + + Yn (Y +)}

+ {(-Yn-) ¾ + (--W ₂₁

+ {(--7 _η ) (7 ₂₁ + Γ ₀ ) -Κ „7 ₁₂ } 7 ₂₂ -Y _n Y ₁₂ (Y ₂ , + Υ ₀ )]

_{/ [{(Υ 22 + (} Υ 21 +)} + (-¾ - 22 + (- - 21 +)]

It is determined using at least one combination that matches Υ, Υ, Υ.

fl f2 f3

[0037] In this case, the measurement result in the reference measurement system can be estimated by correcting the error in the transmission path for the measurement result in the actual measurement system for shunt connection.

[0038] In the second step, at least three of the first correction data acquisition sample or the second correction data acquisition sample are measured using an admittance アド, ,, Υ

mil ml2

When measuring, 、, Y, 測定, ml1 πι21 πι22 m23 is electrically connected between terminal 1 and terminal 2.

May be.

[0039] In another preferred embodiment, in the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample and the second correction data acquisition The sample for use and the electronic component are connected in series. The third step includes pricing in the reference measurement system of at least three samples for obtaining the first correction data having different high frequency characteristics prepared in the first step, and the second step. Acquire at least three second correction data that can be regarded as having the same high frequency characteristics as at least three of the first correction data acquisition samples or the first correction data acquisition samples that have different high frequency characteristics. A sub-step of converting a measured value of the sample for measurement in the actual measurement system into an impedance parameter and deriving a differential impedance component thereof. The mathematical formula is obtained when the electronic component is measured with two ports in the actual measurement system, compared to the impedance when the electronic component is measured in the reference measurement system. Impedance is related through a two-port error model, which has only one port to which two-port differential signals are input that measure impedance when measuring electronic components in the reference measurement system 1 Derived based on the port error model.

[0040] In this case, the 2-port error model is converted by paying attention to the differential impedance component, and the mathematical expression for the 1-port error model is used for high-frequency characteristic error correction. The formula for the 1-port error model can be uniquely determined without considering the sign from the data of the actual measurement system and the reference measurement system of at least three correction data acquisition samples with different high-frequency characteristics. As a result, the correction accuracy is improved, and effects such as mitigation of the influence of noise on the correction accuracy and simplification of the calculation algorithm can be obtained.

[0041] In yet another preferred aspect, in the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample and the second correction data are used. The sample for acquisition and the electronic component are shunt connected. The third step includes pricing the at least three first correction data acquisition samples prepared in the first step, which have different high-frequency characteristics, in the reference measurement system, and the second step. At least three of the second correction data acquisition samples having different high frequency characteristics obtained in the step or at least three of the second correction data acquisition samples having the same high frequency characteristics as the first correction data acquisition sample. A sub-step of converting a measured value of the sample for obtaining correction data in the actual measurement system into an admittance parameter and deriving its in-phase admittance component. The mathematical formula relates the admittance when an electronic component is measured with two ports in the actual measurement system to the admittance when the electronic component is measured with the reference measurement system through a two-port error model. The admittance when the electronic component is measured by the reference measurement system is derived based on a one-port error model having only one port to which the in-phase signal of two ports is input.

[0042] In this case, the 2-port error model is converted by focusing on the in-phase admittance component, and the mathematical expression for the 1-port error model is used for high-frequency characteristic error correction. The formula for the 1-port error model can be uniquely determined without considering the sign from the data of the actual measurement system and the reference measurement system of at least three correction data acquisition samples with different high-frequency characteristics. , Correction accuracy is improved, noise influence on correction accuracy is reduced, and And effects such as simplification of the calculation algorithm can be obtained.

The present invention also provides an electronic component high-frequency characteristic error correction apparatus used in at least the fifth step of the above-described electronic component high-frequency characteristic error correction method. The high-frequency characteristic error correction apparatus for an electronic component uses (a) the mathematical expression determined in step V in the third step and the arbitrary electronic component obtained in the fourth step in the actual measurement system. A storage unit for storing the measured values measured; (b) using the mathematical formula stored in the storage unit to perform an operation for correcting the measurement values stored in the storage unit, And an arithmetic unit that calculates an estimated value of the high-frequency characteristics of the electronic component that would be obtained if measured by the reference measurement system. The invention's effect

[0044] According to the present invention, it is possible to perform the calibration work for the two-terminal impedance component while the measurement system to be corrected remains in the same state as that at the time of actual measurement. As a result, the automatic characteristic sorter, which has not had an effective calibration method, can perform sorting after performing accurate calibration. The user can guarantee the characteristics.

[0045] In addition, the conventional error correction technique requires work that is not in the original measurement, such as removing a terminal from the connector and connecting a standard device for error correction. For this purpose, it is necessary to provide a grounding terminal or to have a structure capable of pressing the short-circuit standard. On the other hand, in the method of the present invention, the measurement for correction may be performed by the same operation as the normal measurement. In addition, there is no need for a GND terminal and a short-circuit mechanism for correction, and the terminal section only needs to have a function that allows normal measurement.

Brief Description of Drawings

[0046] [Fig. L] (a) An explanatory diagram of a measurement system, (b) a circuit diagram of an error model for TRL calibration, and (c) a circuit diagram of an error model for SOLT calibration. (Conventional example)

FIG. 2 is an explanatory diagram of a method for deriving an error factor in SOLT calibration. (Conventional example)

FIG. 3 is an explanatory diagram of a TRL calibration error factor derivation method. (Conventional example)

FIG. 4 is a circuit diagram of an error model for TRRR calibration. (Conventional example)

FIG. 5 is a circuit diagram of an error model for RRRR calibration. (Conventional example) FIG. 6 is an explanatory diagram of measurement positions in TRRR calibration and RRRR calibration. (Conventional example) Sono 7] It is a plan view of a series-connected measurement board. (Conventional example)

Fig. 8 is a plan view of a measurement board for shunt connection. (Conventional example)

[Fig. 9] Fig. 9 is a cross-sectional configuration diagram of a main part showing a configuration of a measurement terminal unit. (Example)

[FIG. 10] (a) Configuration diagram of measurement system, (b) Front view of measurement substrate. (Example 1)

11] It is a graph showing the measurement result of the chip inductor. (Example 1)

(12) (a) Configuration diagram of measurement system, (b) Front view of measurement board. (Example 2)

FIG. 13 is a graph showing measurement results of chip resistance. (Example 2)

14] It is a circuit diagram of an error model for series connection. (Example 1)

15] This is a circuit diagram of the equivalent circuit seen from the port 1 side. (Example 1)

16] It is a circuit diagram of an equivalent circuit seen from the port 1 side. (Example 1)

17] is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 1)

18] It is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 1)

19] It is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 1)

20] It is a circuit diagram of an error model of shunt connection. (Example 2)

FIG. 21 is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 2)

22] It is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 2)

Fig. 23] is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 2)

G. 24] is a circuit diagram of an equivalent circuit viewed from the port 1 side. (Example 2)

25] It is a circuit diagram of an equivalent circuit seen from the port 1 side. (Example 2)

G. 26] is a circuit diagram showing a Z-parameter model of a 2-port circuit. (Examples 3 and 4) Sono 27] FIG. 26 is a circuit diagram showing the T-type equivalent circuit of FIG. (Examples 3 and 4)

FIG. 28 is a circuit diagram showing an equivalent circuit at the time of differential signal input in FIG. 26. (Examples 3 and 4) Sono 29] is a circuit diagram showing a T-type equivalent circuit of the Z parameter of the 2-port error model. (Example 3)

FIG. 30] is a circuit diagram showing an equivalent circuit at the time of differential signal input of FIG. (Example 3) Sono 31] FIG. 30 is a circuit diagram showing an equivalent circuit of FIG. (Example 3)

FIG. 32 is a circuit diagram showing a π-type equivalent circuit. (Example 4) FIG. 33 is a circuit diagram showing an equivalent circuit when an in-phase signal is input in FIG. 31. (Example 4)

FIG. 34 is a circuit diagram showing a π-type equivalent circuit of the Y parameter of the 2-port error model. (Example 4)

FIG. 35 is a circuit diagram showing an equivalent circuit when the in-phase signal in FIG. 33 is input. (Example 4)

FIG. 36 is a circuit diagram showing an equivalent circuit of FIG. 34. (Example 4)

FIG. 37 is a block diagram of a 2-port probe. (Example 3)

Explanation of symbols

[0047] 2 electronic components

20, 21 Measurement board

22a, 22b Transmission line

26 Signal conductor

28 Grounding conductor

BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, embodiments of the present invention will be described with reference to FIGS. 9 to 37.

First, an error correction method for high frequency characteristics of an electronic component according to a first type of embodiment of the present invention will be described with reference to FIGS.

[0050] <Principle 1> The principle of measurement error correction in the case of series connection will be described with reference to FIGS.

[0051] At frequencies above microwaves, the electrical characteristics of electronic components are usually represented by the scattering coefficient matrix. If it is a parameter that can be used, V is easier to use depending on the purpose. As an error parameter when assuming series measurement of a two-terminal impedance element, an impedance T-type connection circuit is used here, and its error model is shown in Fig. 14. In the figure, the part enclosed by a dotted line is the error model for each port. The error model is the terminals 1 and 2 where the object is measured in the reference measurement system, and the correction target.

m m

Is connected between terminals 1 and 2 through which the subject is measured in the measurement system. Variable Z is

d d

Represents impedance dance. The part displayed as DUT is the subject. Since this is a series measurement of 2-terminal impedance elements, the subject is modeled as a 2-terminal impedance element. We believe that the shunt capacity is negligible.

[0052] Observing from port 1, since port 2 is merely a termination impedance, the equivalent circuit of FIG. 15 is obtained. Where Z is the equivalent impedance of port 2.

2

[0053] If FIG. 15 is observed carefully, Z 1, Z 2, and Z are simply a series connection. So Z and Z

If 13 d 2 13 are collectively expressed as Z, the equivalent circuit can be transformed as shown in Fig. 16.

2 el

[0054] Since there are three unknowns Z, Z, and Z in the error model of Fig. 16, acquisition of correction data

11 12 el

These three unknowns can be determined by obtaining three sets of measured values Z when measuring the sample Z for use. Ingredients

d m

Specifically, the three impedance values for the correction data acquisition sample are Z, Z, and Z.

dl d2 d3

If the measured values to be taken are Ζ, Ζ, 関係, the following equation 5a] holds.

mil ml2 ml3

[ ^Equation 5a] f ^flU = z + z ^ el + zd ¹¹

1) Zi _{2 7}

= z + z + z ¹¹

₌ (Z _el + Z, ₃ ) Z ₁₂

m ¹³ ₁₂₊ z _{el +} z, ₃ ¹¹

The error factor can be calculated by the following equation [Equation 5b] obtained from the equation [Equation 5a]. Which solution to choose from among the different soils in the equation will be described later.

[Number 5b]

denom (Z _d2 -Z _dl ) Z _ml3 ― Z _d3 ) Z _ml2 + _-2 [Sat 1 Z _d3 1 Z _d] -J ^ _d3 1 Z _d2 1

-{(Z-Z _d2 ) Z _{m {2} Z _m + {Ζ _ά -Z _d3 ) Z _mU Z _m + (Z _d2 -Z _dX ) Z _m Z _mn )]

I denom

Z _n - Judges _{_{(2 - Z dl) (Z}} j 3 ten _{_{_{_ ~ Z d2 -Z dl Z d3}}} + d2) {^ m mil) m ~ ^ / »ll + ^ mll ^ mil)

I denom

[0056] Z can be obtained by substituting Z and Z in Equation [5b] into Equation [5a] according to the following Equation [5c].

el 11 12

However, it is not used for error correction calculation, that is, Equation [7] described later.

[Equation 5c] ^ n

1 ⁼ z ,, ₊ z ,, -z, ~ ^dl Note that the mathematical formula 5c] can be obtained by using Z and Z instead of Z and Z, or using Z and Z mil dl ml2 d2 ml3 d3.

[0057] Observing from port 2, port 1 is just a termination impedance, so Figure 1

Obtain an equivalent circuit of 7. Where Z is the equivalent impedance of port 1.

[0058] If FIG. 17 is observed carefully, Z 1, Z 2 and Z are simply a series connection. So Z and Z

If 21 d 1 21 are collectively expressed as Z, the equivalent circuit can be transformed as shown in Fig. 18.

1 e2

[0059] Since there are three unknowns in the error model in Fig. 18, Z Z Z, correction data acquisition

21 22 e2

These three unknowns can be determined by obtaining three sets of measured values Z when measuring the sample Z for use.

d m

[0060] Specifically, for each of the three correction data acquisition samples (i = l, 2, 3), assuming that the impedance value is Z and the measured value is Z, the following equation 6a] Is established

di m2i

One.

[Equation 6a] y _ + Ζ (/ ι) 7

f ^{fl21 =} z + z + z ²¹

γ _ + · ¾ ₂ ) 十₂₂ tens γ (Z _e3 + ^ ₃ ) Z ₂₂ I ₇

w ²³ ~ 7 +7 +7 ²¹

[0061] When the error factors Z and Z are obtained from the mathematical expression 6a] for the three correction data acquisition samples (i = l, 2, 3), the following mathematical expression 6b] is obtained. Of the solutions with different soils in the formula,

21 22

Whether to choose one will be described later.

[Equation 6b]

denom (Z _d2 -Z _dl ) Z _m23 + (Z _dX -Z _d3 ) Z _m22 + (Z _{d3 d2} ) ^ _m21

I denom

Z can be obtained by substituting Z and た found in the mathematical formula 6b] into the mathematical formula 6a].

c] The force S required by c], the error correction calculation, that is Absent.

[Equation 6c]

₇ 1 Z _m21 Z ₁₂ -Ζ „Ζ ₁₂

e ² ~ 711 + Z 12-Z m21 ^{d It} should be noted that the formula [Equation 6c] can be obtained by using ZZ or ZZ πι21 dl πι22 d2 m23 d3 instead of ZZ.

[0063] The error model excluding Z and Z is determined as described above.

13 23

[0064] By the way, for Z and Z, just connect the correction data acquisition samples in series.

13 23

Cannot find these values.

[0065] However, since Z and Z are connected in series, it is not necessary to determine their values separately.

13 23

Since there is no error, redraw the error model as shown in Fig. 19. Z in the figure is a series connection of Z and Z (

f 13 23

That is, it is an error factor that can be considered as the sum of values).

[0066] The error model in FIG. 19 is that impedance Z and Z are connected in series between terminal 1 and terminal 1.

m d 11 f

Impedance Z is connected between the connection point of impedance Z and Z and the ground

11 f 12 connected, impedance Z connected between terminal 2 and terminal 2, terminal 2 and ground

d m 21 d

Impedance z is connected between

twenty two

[0067] For example, the impedance seen from port 1 represents the state in which the port 2 side is anti-reflective terminated (that is, normally connected to 50 Ω) in the error model of Fig. 19. Is the force s that can be obtained from the set of the sample value Z for correction data acquisition and the measured value Z when it is connected.

[0068] For the three correction data acquisition samples (i = l, 2, 3), there are three combinations of the correction data acquisition sample value Z and the measurement value Z when this is connected. Formula woman 7] ai mi

The power to calculate z with s. In the formula, Z is the characteristic impedance.

fi 0

[Equation 7] If, =-[{(Z ₂₂ + (Z ₂₁ + Z _Q )) Z _dl + ((Z ₂₁ + Z ₀ ) + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _mU + { (— Z ₁₂ -Z _U ) Z ₂₂ + (-Z ₁₂ -Z _H ) (Z ₂₁

12 21 + ₀ — ZnZ l + ZQ) \

/ [{(Z ₂₂ + (Z ₂₁ + Z ₀ )} Z _mll + (-Z ₁₂ -z _n ) z ₂₂ + (-ζ ₁₂ -ζ _η χζ ₂₁ + ζ ₀ )]

Z / 2 =-[{(+ (Z ₂₁ + Z ₀ )) Z, ₂ + ((Z ₂₁ + Z ₀ ) + Z, ₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _ml2

+ {(— Z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z ₂₁ + Z ₀ )} Z _d2

12 ~

+ Z. )]

/ [{(Z ₂₂ + (Z _{21 +} Z ₀ )} Z _ml2 + (-Z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z) (Z _{21 +} Z ₀ )]

Z Bed _{_{3 = - [{(Z 22}} + (Z 21 + Z 0)) Z rf3 + ((Z 21 + Z 0) + Z 12) Z 22 + Z 12 (Z 21 + Z 0)} Z ml3

+ {(— Z -Z _n ) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z ₂₁ + z ₀ )} z, 3

+ {{—- ^ 12 一_υ ) (· Ζ ₂₁ + ZnZj

/ [{(Z ₂₂ + (Z ₂₁ + Z ₀ )} Z _m + (-Z ₁₂ -Z „) Z ₂₂ + (-Z ₁₂ -Z„) (Z ₂₁ + Z ₀ )] [0069] Z Since there is only one value, Z,, Ζ obtained in the formula 7] should have the same value.

1 fl f2 f3

As shown in the formulas 5b] and 6b], 力, Ζ, Ζ, Ζ have different signs.

12 21 21 22

There are two solutions, and depending on the combination, Ζ, ない, Ζ will not work.

fl f2 f3

[0070] Therefore, for each of 2 ⁴ = 16 combination patterns shown in Table 1 below, Z, Z, Z of Z, Z, Z are calculated by calculating ZZ, Z of the above-mentioned mathematical formula 7 , Z

fl f2 f3, fl f2 f3 12 21 21 22 The combination is selected. There are multiple combinations that match Z and Z Z

fl f2 f3

Therefore, any one of them may be used.

1] Code z ₂₁ code z ₁₂ code Z22 code

/ 《Turn 1 + + + +

Pattern 2 + + + ―

Pattern 3 + + one +

Pattern 4 + + 1 ―

Pattern 5 + one + +

Pattern 6 + ― + One

Pattern 7 + one-+

Pattern 8 +-One ―

Pattern 9 ― + + +

Pattern 10-+ + one

Pattern 11 one + one +

Pattern 12-+ one

Pattern 13 ―-+ +

Pattern 14 1 +-Pattern 15 ―-― +

Pattern 16 ― One ― One [0071] Since Z and Z are error factors that form Z as a series connection,

13 23 f

As long as the correction data acquisition sample is a series connection of two-terminal impedance elements, if the correction is performed based on the error model of FIG. 19, the same result as the correction based on FIG. 14 can be obtained.

<Principle 2> The principle of 2-port error correction during shunt measurement will be described with reference to FIGS.

[0073] When a shunt measurement of a two-terminal impedance element is assumed, an impedance π-type connection circuit (which is not very common as a circuit parameter) is adopted as an error parameter. Figure 20 shows. In the figure, the part enclosed by the dotted line is the error model of each port. The error model is the terminals 1 and 2 where the object is measured by the reference measurement system and the object by the measurement system to be corrected. Mmdd between terminals 1 and 2

Connected. The variable represents admittance. The circuit model is different from that of series measurement, but they can be converted to each other. The part displayed as DUT is the subject. Since it is a shunt measurement of a two-terminal impedance element, the subject can be modeled as a two-terminal impedance element.

[0074] As in the case of series measurement, the purpose of correction is to derive the value of the error model parameter in the figure from the measurement result of the correction data acquisition sample. Again, the correction data acquisition sample should be connected only in the state shown in the figure, so that the measurement jig is complicated and does not cause any problems!

[0075] Now, although the equivalent circuit is different at first glance, the error model parameters can be determined by almost the same procedure as in the case of series connection as follows.

First, when observing from port 1, port 2 is merely a terminal admittance, so the equivalent circuit of FIG. 21 is obtained. Where Y is the equivalent admittance of port 2.

2

[0077] Since Y, Y, and Y in Fig. 21 are in a parallel connection relationship, Y and Y are collectively displayed as Y.

13 d 2 13 2 el Then, the equivalent circuit can be modified as shown in FIG.

[0078] Since the number of unknowns in the error model in Fig. 22 is three as in the case of series measurement, these unknowns can be determined by measuring three correction data acquisition samples. If the variable name is determined following the series measurement, the following mathematical formula 8a] holds. γ _ (l +) I γ

+

γ, + D 2! V

½ '1 2

γ _ V ^J el + 3 2 γ

2 ^{+ 1} e \ ⁺ 'di

[0079] The error factor can be calculated by the following mathematical formula 8b] obtained from the mathematical formula 8a].

[Equation 8b]

denom = (Y _d2 -Y _dl ) Y _mU + (-Y _d ,) Y _mll + (-Y _d2 ) Y _mll

) 11 = Sat [Satichi] 1 1 1 H

-K-) 2 13 +, 1-3) m〗 3 + (-〃 ml mu}]

I aenom

_{Y 12 = ± (-) (} + (- -) ^ +) (ml2 - (3 + (- 2 -r mll) r mI3 + r mll ^ I2)

I denom

[0080] Y can be calculated by substituting Υ Υ found in Equation [Equation 8b] into Equation [Equation 8a].

el 11 12

It is not used in the calculation of force S and error correction calculated by 8c], that is, the mathematical formula 10 described later.

[Equation 8c] γ _ γ

数式 Note that the mathematical formula [Equation 8c] can be obtained by using Υ Υ instead of Υ Υ, or Υ Υ

The power can be obtained using mil dl ml2 d2 ml3 d.

Three

[0081] Actually, this mathematical expression 8b] is substantially the same mathematical expression as in the case of series measurement. Which solution to choose from among the different soils in the equation will be described later.

Next, the derivation of unknowns will be described when viewed from port 2! /.

[0083] When observing from port 2, since port 2 is merely a terminal admittance, the equivalent circuit of Fig. 23 is obtained. Where Y is the equivalent admittance of port 1.

[0084] Since Υ, ,, and Y in Fig. 23 are connected in parallel, Y and Y are collectively displayed as Y.

23 d 1 23 1 e2 Then, the equivalent circuit can be modified as shown in FIG.

[0085] If the variable name is determined in the same way as for port 1, the error factor can be calculated in the same way, and the following formula 9a] holds.

9a] γ _ ('el ⁺ ^ d \) ^ 22 ._ γ

^Im21 ~ ¹ ^v22 + ^ γ ¹ e2 + Τγ 1 ²¹

γ _ +).

^Im22 ~ γ + + ²¹

22 `` ¹ ί 2

_ ( _{1 +} r _d3 ) r ₂₂

^J 22 "J el 卞" * </ 3

The error factor can be calculated by the following mathematical formula 9b] obtained from the mathematical formula 9a].

[ _Equation 9b] denom = (Y _d2 -Y _dl ) Y _m23 + (Y _d -Y _d3 ) Y _ml2 + (-Y _d2 ) Y _m21

^ 21 ~ [Sat Λ / 2-1]-1--1-m23

-{(-2)] w22 ^ n23 + (1-) 21 23 + (-) 2 m2]

I denom

Y ₂₂

I denom

[0087] Y can be obtained by substituting Y and Y obtained by the formula [Equation 9b] into the formula [Equation 9a].

However, it is not used in the error correction calculation, that is, the mathematical formula 10 described later.

[Equation 9c]

V V-V V

γ _ ¹ m21 ^J 12 -Ί1-Ί2 γ Note that the equation [Equation 9c] uses Y, に instead of Y, Υ, or Υ, ι πι21 dl πι22 d2 m23 d, The power to seek s.

Three

[0088] Y and Y, which are error factors that have not yet been obtained by the above procedure, are used for obtaining correction data.

13 23

Since it is a force-parallel connection relationship that cannot be obtained simply by connecting the sample to the shunt, it is not necessary to set the values separately, so the error model is drawn as shown in Fig. 25. Revisit. Y in the figure is an error factor that can be thought of as parallel connection of Y and Υ (that is, the sum of values).

f 13 23

is there.

[0089] In the error model of FIG. 25, admittance Y is connected between terminal 1 and terminal 1, and terminal 1

m d 12

Admittance Y is connected between the connection point of admittance Y and ground and admittance m 12 11

The admittance Y is connected between the connection point between the terminal Y and the terminal 1 and the ground, and the terminal 2 and the terminal

12 d f d

2 is connected to admittance Y, and the connection point between admittance Υ and terminal 2 and ground m 22 22 m

Is connected to admittance γ.

twenty one

[0090] For example, the impedance viewed from port 1 represents the state in which the port 2 side is anti-reflective terminated (that is, normally connected to 50 Ω) in the error model of Fig. 23. Can be obtained from the set of the correction data acquisition sample value Υ and the measurement value 際 when it is connected. This is also the same as in the case of series measurement, and it is possible to calculate the following mathematical formula [10] -CY. Y in the formula indicates the characteristic admittance.

γ =-[{(¾ + (7 ₂₁ + Y ₀ w _dl + ((r ₂ , + Y _Q ) + Y _U ) Y ₂₂ + „+ {(-¾-r ₁₁ ) (7 ₂₁ +7 ₀ ) -^ ₁ ¾} r ₂₂ -7 ₁₁ ¾ (7 _{21 +} r ₀ )]

/ [{(¾ + (¾ +)} + (- - 22 + (-Y -YnW + Yo)] γ ί2 = ya _{_{2 + σ 2, + Υο)}} ) + (+ γ 0) + y l2) ¾ + (^ 2, +)} ₂

+ {(--Y _U ) Y ₂₂ + (-Y ₁₂ -7 _n ) (7 ₂₁ + r ₀ )} r _rf2

+ {(--) + ₀ ) -7 _U F ₁₂ } ¾ -YuYniYn + ^ o)]

/ [{(¾ ₂ + (-¾—) ¾ + (-H ₂₁ +)]

Y _f3 =-[{( ₂₂ + (7 ₂₁ + Y ₀ )) + (( ₂₁ + Γ ₀ ) + Υ _η ) Υ ₂₂ tens Y _i2 (Y ₂₁ + Υ _α )} 7 _ml3

+ {(-Γ ₁₂ -Y) Y _{n +} (-Y _n -7 _U ) (7 ₂₁ + Y ₀ )} Y _d3

+ {(-Yn-,) (¾ + Yo)-YuYn ¾ + ^ ο)]

/ [{(¾ + (Y _2l + Y _a } Y _m u-22 + (~ ^Υ η H +]

[0091] Since there is only one value of Υ, Υ, Υ, and 求め obtained in Equation 10 should have the same value

f fl f2 f3

Y, Υ, Υ, Υ are given by the sign as shown in force S, formula woman 8b] and woman 9b]

12 12 21 22

There are two different solutions, and depending on the combination, Υ, Υ, Υ will not work.

fl f2 f3

[0092] Therefore, for each of the 2 ⁴ = 16 combinations shown in Table 2 below, Y, Y, Y of the above formula 10 is calculated, and Y, Y, Y matches Y, Y , Y, Let's choose the combination of Y.組み合わせ Υ, Υ matches more than one combination

22 fl f2 f3

Since there are several, any of them may be used.

[Table 2]

[0093] Y and Y are error factors that form Y by connecting them in parallel.

13 23 f

As long as the material is a shunt connection of a two-terminal impedance element, if the correction is performed based on the error model in FIG. 25, the same result as the correction based on FIG. 20 can be obtained.

Next, examples will be described with reference to FIGS. 10 to 13.

<Example 1> A case of series connection will be described with reference to FIG. 10 and FIG. Series connection is a method of connecting a device under test between two ports of a measuring machine.

[0096] In the measurement system to be corrected, as shown in the overall configuration diagram of Fig. 10 (a) and the front view of the measurement board 20 in (b), the electronic component 2 as the subject is placed on the measurement board 20. It is arranged so as to straddle the slit 22x between the transmission lines 22a and 22b formed on the upper surface, and is connected in series between the transmission lines 22a and 22b. SMA connectors 56 66 are soldered to both ends of the transmission lines 22a, 22b; 24 on the upper and lower surfaces of the measurement board 20, and are connected to the network analyzer 70 via coaxial cables 58, 68. The network analyzer 70 uses an Agilent network analyzer 8753D, and the measurement board 20 is designed with a characteristic impedance of 50 Ω. The measurement board 20 has a length L of 50 mm and a width W of 30 mm.

In the reference measurement system, measurement is performed with an Agilent measurement jig 16192A attached to an Agilent impedance analyzer 4291. [0098] The electronic component 2 that is the subject is a 56 nH chip inductor of 1.0 mm X 0.5 mm size.

[0099] Measurement and correction operations will be described in order.

[0100] (1) Prepare three correction data acquisition samples. The three correction data acquisition samples include

Resistors of 2.2 Ω, 51 Ω, and 510 Ω were used.

[0101] (2) Measure the impedance Ζ, Ζ, の of the sample for acquiring correction data using the reference measurement system.

dl d2 d3

The Note that the number of measurement points and the sweep frequency range must be standardized by the reference measuring instrument and the network analyzer actually used.

[0102] (3) Calibrate the transmission line to the end of the coaxial cable in the measuring instrument (8753D) actually used for measurement. This calibration can be performed by a generally performed SOLT calibration.

[0103] (4) Measuring instrument (8753) that actually uses the impedance of the correction data acquisition sample for measurement

Measure in D). At that time, Z, mi z, z and z, z, z are acquired in the same number of measurement points and sweep frequency range as the reference measuring machine.

ml2 ml3 m21 m22 m23

[0104] (5) The correction coefficient is calculated from the measurement data of the reference measuring instrument (4291) and the measuring instrument (8753D) actually used for measurement on a personal computer based on the principle 1> described above. The procedure up to this point is the measurement system correction procedure.

[6] (6) Measure the chip inductor with the measuring instrument (8753D) that is actually used for the measurement.

[0106] (7) Using the measurement data and correction data, the corrected measurement is calculated by a personal computer.

[0107] As a result of performing the measurement and correction process according to the above procedure, the measurement result of the reference measuring instrument and the measurement of the network analyzer matched.

FIG. 11 is a graph showing the results of measurement and correction processing performed on a 1005 size chip inductor (52 nH). Figure 11 (a) is a graph of the reference value, the measurement value before correction, and the measurement value after correction. The “reference value” is a value measured with a reference measuring machine. “Before correction” is the measurement result itself with the measuring instrument that is actually used for measurement, and is corrected, measured, and measured value. “After correction” is a value obtained by correcting the measured value of the measuring instrument actually used for measurement (estimated value of the measured value when measured with the reference measuring instrument). Fig. 1 l (b-1) is a graph of measured values before correction, Fig. 1 l (b-2) is a graph of measured values after correction, and Fig. 11 (c) is a graph of reference values. It is. [0109] As shown in Fig. 11 (a), "reference value" and "after correction" agree well enough that they cannot be distinguished in the figure, but "before correction" is different from "reference value". There is a big shift. In other words, if correction is not performed, only measurement values that are significantly different from those measured with the reference measurement device can be obtained, but by performing correction, measurement values that are very close to the measurement values obtained with the reference measurement device can be obtained. Power S can be.

<Embodiment 2> A case of shunt connection will be described with reference to FIG. 12 and FIG. Shunt connection is a method of connecting a device under test between one port of the measuring instrument and the ground.

[0111] In the measurement system to be corrected, the electronic component 2 that is the subject is placed on the upper surface of the measurement substrate 21, as shown in the overall configuration diagram of FIG. It is connected between the formed signal conductor 24 and the ground conductor 25. The measurement board 21 has SMA connectors 56 and 66 soldered to both ends of the signal conductor 24 and the ground conductor 25, and is connected to the network analyzer 70 via coaxial cables 58 and 68. A network analyzer 8753D manufactured by Agilent is used as the network analyzer 70, and the measurement board 20 is designed with a characteristic impedance of 50Ω. The measurement board 20 has a length L of 50 mm and a width W of 30 mm.

[0112] The reference measurement system is the Agilent impedance analyzer 4291 attached to the Agilent measurement jig 16192A and measured.

[0113] The electronic component 2 which is the subject is a 50 Ω chip resistor of l.Omm × 0.5mm size.

Next, the measurement and correction operations will be described in order.

[0115] (1) Prepare three correction data acquisition samples. Resistors of 2 · 2 Ω, 51 Ω, and 510 Ω were used.

[0116] (2) Measure admittance Υ, Υ, Υ of the sample for acquiring correction data using a reference measuring instrument.

dl d2 d3

Note that the number of measurement points and the sweep frequency range must be standardized for the reference measuring instrument and the network analyzer used.

[0117] (3) Calibrate the transmission line to the end of the coaxial cable in the measuring instrument (8753D) actually used for measurement. This calibration can be performed by a generally performed SOLT calibration.

[0118] (4) Measure the admittance of the correction data acquisition sample with the measuring instrument (8753D) that is actually used for the measurement. At that time, Y, Y mi l in the same number of measurement points and sweep frequency range as the reference measuring machine

, Y and Y, Y, Y are obtained. [0119] (5) The correction coefficient is calculated from the measurement data of the reference measuring instrument (4291) and the measuring instrument (8753D) actually used for measurement on a personal computer based on the principle 2> described above. The procedure up to this point is the correction procedure.

[6] (6) Measure the chip resistance with a measuring instrument (8753D) that is actually used for measurement.

[0121] (7) Using the measurement data and correction data, calculate the corrected measurement using a personal computer.

[0122] As a result of performing the measurement and correction process according to the above procedure, the measurement result of the reference measuring instrument and the measurement of the network analyzer matched.

FIG. 13 is a graph showing the results of measurement and correction processing performed on a 1005 chip resistance (50 Ω). Figure 13 (a) is a graph of the reference value, the measurement value before correction, and the measurement value after correction. The “reference value” is a value measured with a reference measuring machine. “Before correction” is the measurement result of the measuring instrument that is actually used for measurement, and is the measurement value that has not been corrected. “After correction” is a value obtained by correcting the measured value of the measuring instrument actually used for the measurement (estimated value of the measured value when measured with the reference measuring instrument). Fig. 13 (b-1) is a graph of measured values before correction, Fig. 13 (b-2) is a graph of measured values after correction, and Fig. 13 (c) is a graph of reference values. .

[0124] As shown in Fig. 13 (a), "reference value" and "after correction" match well enough to be indistinguishable from the figure, but "before correction" is different from "reference value". There is a big shift. In other words, if correction is not performed, only measurement values that are significantly different from the measurement values obtained with the reference measurement device can be obtained. You can get the power S.

[0125] In the first type embodiment of the present invention described above, in a two-port measurement system in which a two-terminal impedance element is connected in series or shunted to a measurement jig or probe, the electrical characteristics of each port are measured. Is represented by a T-type (Fig. 14) and π-type (Fig. 20) equivalent circuit, which is usually modeled with six errors in a reversible circuit, but is simplified to five errors. By doing so, when deriving the error of the measurement jig or probe, three standard two-terminal impedance elements (hereinafter referred to as standard samples) are used, and the standard sample is grounded in series connection. Without the shunt connection, five error values can be derived without cutting the signal line.

[0126] When the above error model is used, a measurement jig derived from the measured values of three standard samples Of each error of the probe, four have two solutions with different signs. Therefore, to confirm which four error sign combinations are correct, confirm that the three values derived from each of the three standard samples are the same for the remaining one error in each sign combination. It is decided by going.

[0127] However, it is difficult to determine which combination of codes is correct due to measurement variations of the three standard samples and trace noise of the measuring instrument. There is no choice but to do it. For this reason, incorrect code combinations may be selected due to measurement variations and noise, and there may be errors and frequencies that cannot ensure correction accuracy.

Next, an error correction method for high frequency characteristics of an electronic component according to a second type of embodiment of the present invention will be described with reference to FIGS. 26 to 37. FIG.

[0129] In the second type embodiment of the present invention, by using a one-port error model equivalent to the two-port error model, the combination of codes as in the first type embodiment of the present invention can be realized. No selection is required. This improves the accuracy of deriving error correction parameters and the accuracy of measurement error correction. Details will be described below.

<Balance Conversion 1-Port Error Model> First, the balance conversion 1-port error model used in the second type of embodiment of the present invention will be described. Balance conversion The 1-port error model is obtained by performing the balance conversion on the 2-port error model as follows.

FIG. 26 shows a circuit diagram of a model in which a 2-port circuit is represented by Z parameters. The relationship shown in Fig. 26 can be expressed as a determinant as shown in the following equation 11].

[Equation 11]

(νΛ (ίΙτ _λヽノ

V 2 J

[0132] Replace the input value with V = (V + V) / 2, I = (I +1) / 2, and change the output value to V = (V—V

C 1 2 C 1 2 d 1 2

) / 2, I = (I -I) By performing a balance transformation that replaces / 2, the Z parameter is d 1 2

It is converted as follows: (νΛ (z _cc ^, ヽ

[0133] For the Y parameter, the following number

+ 1

J 21 ¹

[0134] Since the reversible theorem holds in ordinary passive circuits, Z = Z, Y if Z parameters

12 21 If the parameter, then Υ = Υ. From there, the Ζ parameter and the Υ parameter are

12 21

And a π-type equivalent circuit.

[0135] When the の parameter shown in Fig. 26, Equation Woman 11] is converted into a 等価 -type equivalent circuit, the circuit diagram in Fig. 27 is obtained. Fig. 27 is transformed into an equivalent circuit for differential signal input, resulting in the circuit diagram of Fig. 28. The direct IJ impedance component (Ζ — Ζ) + of ports 1 and 2 in the vertical equivalent circuit +

11 12

(Ζ Ζ), it can be seen that it is における in the mathematical formula 12].

22 12 dd

By utilizing this, when the two-terminal impedance element is connected in series, the correction model is as follows.

[0137] FIG. 29 is a circuit diagram showing a two-port error model using a T-type equivalent circuit when two-terminal impedance elements are connected in series using a measurement jig. The 2-port error model is the part enclosed by the dotted line, and when measuring electronic components with the reference measurement system. Impedance Z and impedance when measuring electronic components with the actual measurement system. It is connected between two specified ports (Portl, Port2).

As in the case of FIG. 28, when the circuit of FIG. 29 is transformed into an equivalent circuit when a differential signal is input, the circuit diagram shown in FIG. 30 is obtained.

[0139] The series impedance between ports 1 and 2 in the circuit of Fig. 30 is calculated as follows:

14].

[Equation 14]

Z '= Ze-Ze + "e-Ze' ( ^Zg 22- ^Ze i2 + ^ζ -Ze ^ + _d ) x (Ze ₁₂ + Ze ^)

[0140] This mathematical model 14] is equivalent to the differential impedance component Zt of the Z parameter measured by connecting two-terminal impedance elements in series using a measurement jig. That is, the formula [

dd

Equation 14] is an actual measurement for impedance Z when an electronic component is measured with a reference measurement system.

d

Associate the impedance when measuring electronic components with two ports in the measurement system via a two-port error model. Equation 14] shows that the circuit of FIG. 30 is equivalent to the circuit shown in FIG.

[0141] In Fig. 31, the impedance components in the circuit are combined into three, which is the same as the one-port error model in which the error of the measurement tool is represented by a T-type equivalent circuit. This indicates that the impedance of the DUT can be derived by performing measurement and correction in the following steps (1) to (4) in the case of series connection!

[0142] (1) Measurement jigs for three correction samples (chip resistance, etc.) whose characteristics (impedance) are priced, or three correction samples that can be considered to have high frequency characteristics equivalent to these three correction samples Use to measure the Z parameter. A network or impedance analyzer is used for the measurement.

[0143] (2) Perform Z parameter balance conversion and extract the differential impedance component Z

dd

Yes

[0144] (3) The three error components of the measurement jig, summarized as the circuit in Fig. 30, which is an equivalent circuit of the differential impedance component, are the Z when the three correction samples are measured and the "three Correction test

dd

Calculated from the priced characteristics (impedance) of the material.

[0145] (4) Remove the three error components derived in the procedure of (3) from Z when measuring the DUT. By doing this, calculate the impedance Z of the DUT.

Note that the error Z parameters of the measurement jig of FIG. 29 are summarized as shown in FIG. 31, so that each value cannot be obtained independently, but there is no problem in performing the correction.

[0147] When a two-terminal impedance element is shunt-connected, the correction model is as follows

Yes

[0148] For shunt connection of two-terminal impedance elements, by converting the Y parameter to a π-type equivalent circuit and using an in-phase admittance component, it can be used as a model for 1-port correction as in series connection. .

[0149] The circuit diagram of Fig. 32 shows a circuit diagram of a π-type equivalent circuit using the Υ parameter. If Fig. 32 is transformed into an equivalent circuit when a common-mode signal is input, the parallel admittance component of ports 1 and 2 in the π-type equivalent circuit is Υ in Equation 13] as shown in the circuit diagram of Fig. 33. I understand.

[0150] Next, as with the series connection, the 2-port error model in the case where a 2-terminal impedance element is shunt-connected to the circuit diagram of Fig. 34 using a measurement jig is measured as a π-type equivalent. This is shown using a circuit. The 2-port error model is the part enclosed by the dotted line. Two ports (the admittance とき when measuring electronic components with the reference measurement system and the admittance when measuring electronic components with the actual measurement system) Portl and Port2). As before, if the circuit in Fig. 34 is transformed into an equivalent circuit when an in-phase signal is input, the circuit diagram in Fig. 35 is obtained.

[0151] When the parallel impedance between ports 1 and 2 in the circuit of FIG. 35 is obtained, the following mathematical formula 15] is obtained.

[Equation 15]

Yt _{d ά} = Ye _n -Ye ₁₂ + Ye ^ _-Ye _M + ^ —— ^ —— -——-~ — ~ ———— d Ye ₂₂ -Ye _l2 + Ye ₃₃ -Ye _M + Y _d + Ye ₁₂ + Ye ₃₄

[0152] Equation 15 is equivalent to the in-phase admittance component Yt of the Y parameter measured by shunting a two-terminal impedance element using a measurement jig. In other words, the mathematical formula 15]

Admittance when measuring electronic components with two ports in the actual measurement system versus admittance Y when measuring electronic components with the reference measurement system, and a 2-port error model Associate through. Equation [15] shows that the circuit of FIG. 34 is equivalent to the circuit shown in FIG.

In FIG. 35, there are three admittance components in the circuit, which is the same as the one-port error model in which the error of the measurement jig is represented by a π-type equivalent circuit. This is because, as with the series-connected vertical type equivalent circuit, the balance conversion of the vertical parameter measured using the measurement jig is performed, and then 1-port correction is performed for the common-mode admittance component. Demonstrate that DUT admittance can be derived! /

<Example 3> A measurement error correction procedure when a two-terminal impedance element is connected in series in an actual measurement system will be described.

[0155] Three types of two-terminal two-port devices (appropriate chip resistors, devices themselves, etc. can be used) are priced by the TRL calibration method and the RR RR calibration method for impedance analyzers and network analyzers (hereinafter referred to as the following) Standard 2-port device). This is a procedure that should be done on the desk. This pricing is a pricing in the reference measurement system.

[0156] Next V, connect the standard 2-port sample or a sample that can be regarded as having high-frequency characteristics equivalent to the standard 2-port sample to the two ports of the actual measurement system, and measure the S parameter.

[0157] Next, the above S-parameter measurement result is converted into a differential Ζ parameter using the following equation 16). This mathematical model 16] is derived by converting the Ζ parameter on the right side of the mathematical model 14] described above into an S parameter.

[Equation 16] _ 2 X ₀ X

-

_{_{^ Π + S M11 + S M11}} x S Mll - S MU x S M12 - 1

Where, S: Actual S-parameter measurement result in 2-port measurement system

M

Z: Converted differential Z parameter

dd

z: Characteristic impedance of measurement system

0

[0158] Next! /, Represents the relationship between the priced value of the standard 2-port sample and the converted differential Z parameter using the 1-port error model. The 1-port error model is There is no problem even if the relationship is modeled by converting to a reflection coefficient.

[0159] Next, the error parameter of the 1-port error model is calculated from the relationship between the impedance values determined for the three standard 2-port samples and the converted differential Z parameter. When using the one-port error model shown in Figure 31, the three unknowns shown in Figure 31 as error parameters, namely Ze — Ze + Ze — Ze and Ze — Ze + Ze — Ze.

11 12 44 34 22 12 33 34, Ze +

12

Calculate Ze.

34

[0160] Next, using the 1-port error model that uses the calculated error parameter values, the values measured by the actual measurement system for the other 2-terminal 2-port devices are corrected, and the true values of the other 2-terminal 2-port devices are corrected. Obtain a value (ie, an estimate of the measurement that would be obtained if measured with a reference measurement system).

[0161] As described above, by using balance conversion for the measured 2-port measurement system and replacing the relationship with the priced value of the standard 2-port sample with the 1-port error model, the error parameter is uniquely determined. Derived.

[0162] According to this method, as in the AAA correction method (Example 1), the error parameter is not uniquely determined, and no problem arises. The effects of variations and measuring instrument trace noise are alleviated.

[0163] Therefore, the error parameter derivation accuracy and the correction accuracy are improved by the AAA correction method (Example 1).

<Example 4> A measurement error correction procedure when a two-terminal impedance element is shunt-connected in an actual measurement system will be described.

[0165] First, three types of two-terminal two-port devices (appropriate chip resistors, devices themselves, etc.) can be priced by the TRL calibration method or RRRR calibration method for impedance analyzers and network analyzers (hereinafter referred to as the following) Standard 2-port device). This is the procedure that should be done on the desk. This pricing is a pricing in the reference measurement system.

[0166] Next, the standard 2-port sample is connected in the 2-port actual measurement system, and the S parameter is measured.

[0167] Next, the S parameter measurement result is converted into an in-phase Y parameter using the following mathematical model 17]. This mathematical model 17] uses the S parameter for the Y parameter on the right side of the mathematical model 15] described above. Derived by converting to a parameter.

Where SM: Actual S-parameter measurement result in 2-port measurement system

Y: converted in-phase Y parameter

cc

z: Characteristic impedance of measurement system

0

[0168] The relationship between the priced value of the standard 2-port sample and the converted in-phase Y parameter is expressed using a 1-port error model. The one-port error model can be converted into a reflection coefficient instead of the one shown in Fig. 36, and the relationship can be modeled.

[0169] Next, the error parameter of the 1-port error model is calculated from the relationship between the admittance values of the three standard 2-port samples and the converted in-phase Y parameter. When using the one-port error model in Figure 36, the three unknowns shown in Figure 36 as error parameters: Ye + Ye, Ye — Ye + Ye Ye, Ye — Ye + Ye Y

12 34 11 12 44 34 22 12 33 e is calculated.

44

[0170] Next, using the 1-port error model that uses the calculated error parameter values, the values measured by the actual measurement system for the other 2-terminal 2-port devices are corrected, and the true values of the other 2-terminal 2-port devices are corrected. Obtain a value (ie, an estimate of the measurement that would be obtained if measured with a reference measurement system).

[0171] As described above, the error parameters are uniquely derived by replacing the relationship with the priced values of the standard 2-port sample with the 1-port error model using balance conversion for the actual 2-port measurement system. Is done.

[0172] According to this method, the error parameter is not uniquely determined as in the AAA correction method (Example 2), and no problem arises. The effects of variations and measuring instrument trace noise are alleviated.

[0173] Therefore, the error parameter derivation accuracy and the correction accuracy are improved by the AAA correction method (Example 2). <Specific example> The configuration diagram of Fig. 37 shows the structure of the 2-port probe used for the measurement. Other experimental conditions are as follows. The connection method is series connection.

[DUT] Correction type confirmation 1608 size chip resistor 2 types (100 Ω, 392 Ω)

[Measuring instrument] Ε8364 (Agilent Technologies)

[Measurement frequency] 50MHz ~; 1. 8GHz

[Measurement points] 801 points

[Intermediate frequency] 300Hz

[Correction sample] 3 types of 1608 size resistors (1.2 Ω, 47 Ω, 560 Ω) priced by impedance analyzer 4291 A (Agilent Technologies)

[0175] Table 3 shows the results of correcting the chip resistance (100 Ω, 392 Ω) using the method of Example 3 and the AAA correction (Example 1) using the 1-port error model in Fig. 31. .

[Table 3] Comparison of correction errors

[0176] Table 3 shows the average correction error and 3σ. Except for the average value of the resistance of 392 Ω, which is marked with “*” to the value with the smallest correction error, all values show the values with the smallest correction error. Also, in AAA correction method, measured value S of port 1 and port

11

Measured value of 2 S force Each of the calculated correction values differs from the average value of the correction errors.

twenty two

It is possible to read even if there is. Therefore, Table 3 shows that the method of Example 3 can be corrected more accurately than the AAA correction method (Example 1).

<Summary> By using the error correction method described above, it is possible to perform the calibration work for the two-terminal impedance component while the measurement system to be corrected remains in the same state as at the time of actual measurement. For this reason, it is possible to perform error correction of the measurement system even if the mass production device itself or a sample with approximately the same size and shape as the mass production device can be connected to the measurement terminal V it can. [0178] It should be noted that the present invention is not limited to the above-described embodiment, and can be implemented with various modifications.

For example, the present invention can be applied not only to a measurement system using a measurement substrate but also to a measurement system using a measurement pin.

Claims

The scope of the claims

[1] From the results of measuring an electronic component, which is a two-terminal impedance component, using an actual measurement system, calculate an estimate of the high-frequency characteristics of the electronic component that would be obtained if the electronic component was measured using a reference measurement system. An electronic component high frequency characteristic error correction method comprising:

First step of preparing at least three first correction data acquisition samples that are priced in the reference measurement system and have different high-frequency characteristics;

At least three of the first correction data acquisition samples or at least three second correction data acquisition samples that can be regarded as having the same high-frequency characteristics as the first correction data acquisition sample A second step to measure with

Pricing data in the reference measurement system of the first correction data acquisition sample prepared in the first step and the first correction data measured in the actual measurement system in the second step From the measurement data of the acquisition sample or the second correction data acquisition sample, the measurement value measured by the actual measurement system and the measurement value measured by the reference measurement system are used as the error correction coefficient of the transmission path. A third step of determining a mathematical expression to be used, a fourth step of measuring an arbitrary electronic component with the actual measurement system,

Based on the measurement result obtained in the fourth step, it can be obtained if the electronic component is measured by the reference measurement system using the mathematical formula determined in the third step. A fifth step of calculating an estimated value of the high-frequency characteristic of the electronic component, and a method for correcting the high-frequency characteristic error of the electronic component, comprising:

[2] In the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample, the second correction data acquisition sample, and the electronic component are Connected, series,

The mathematical formula is expressed in terms of terminals 1 and 2 where impedance Z is measured when an electronic component is measured in the reference measurement system, and impedance m m when the electronic component is measured in the actual measurement system.

Derived based on an error model connected between terminals 1 and 2 where Z is measured, d d d

When calculating the impedance of the terminal 1 force, the error model is the terminal m

Impedances z and z are connected in series between 1 and the terminal 1, and the impeder m d 11 f

Impedance Z is connected between the connection point of the resistors Z and Z and the ground, and the terminal 2 Is connected between the terminal 2 and the ground.

twenty one

Impedance ζ is connected,

twenty two

The impedance Ζ Ζ Ζ Ζ Ζ is

f 11 12 21 22

As a result of measuring the impedance of at least three of the first correction data acquisition samples in the first step, Z 1, Z 2, and Z,

dl d2 d3

In the second step, the impedance of the terminal 1 is measured for at least three of the first correction data acquisition sample or the second correction data acquisition sample.

m

Measurement result Z 1, Z 2, Z 3 and impedance of the terminal 2

mil ml2 ml3 m m21

, z z and

m22 m23

The following formula [number la]

[Number la]

denom (Z _d2 -Z _d ) Z _mU + (Z _dl Z _d3 ) Z _ml2 + (Z _d3 Z _d2 ) Z _m

― [

/ denom

I denom The following formula [number lb]

[Several lb]

_{_{denom = (Z d2 -Z dl)}} Z m23 + {Z d -Z d) Z mn + (Z ^ 3 -Z d2) Z m2l L mechanic one ^A one "\ / Z one Z _m2 i

One _{_{_{2) ZZ d3) Z m2l Z}}} m23 J

Of the 16 Z Z Z Z stitches obtained from I denom I denom,

11 12 21 22

For the following formula [Equation 2]

[Equation 2] Z _fl =-[{(Z ₂₂ + (Z ₂₁ + Z ₀ )) Z _dl + ((Z ₂₁ + Z + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _m

+ {(-Z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z _U ) (Z ₂ , + Ζ ₀ )} Ζ _Λ

Z _U ) (Z ₂₁ + Ζ ₀ ) -Ζ _Π ^ _Ι2 } Ζ _{22 U} Z ₁₂

/ [{(+ (Z ₂₁ + Z ₀ )} Z _mll + (— Z _l2 -Z „) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z ₂₁ + Z ₀ )]

Z _{/ 2} =-[{(Z ₂₂ + (Z ₂₁ + Z ₀ )) Z _d2 + ((Z ₂₁ + Z ₀ ) + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _ml2

+ {(—Z ₁₂ -Z _U ) Z ₂₂ + (-Z ₁₂ -Z _n ) (Z ₂₁ + Z ₀ )} Z _d2

+ ((— Z ₁₂ — Z _u ) Z ₀ ) —, _u Z _n Z ₂₂ 1 Z _U Z ₁₂ (Z ₂₁ + Z ₀ ) J

/ [{(Z ₂₂ + (Z ₂₁ + Z ₀ )} Z _ml2 + (-Z ₁₂ -Z _n ) Z ₂₂ + (-z ₁₂

+ Ζο)] z ₃ = 1 [{(+ (Z ₂₁ + Z ₀ )) Z _rf3 + ((Z ₂₁ + Z ₀ ) + Z ₁₂ ) Z ₂₂ + Z ₁₂ (Z ₂₁ + Z ₀ )} Z _ml3

+ {(-Z, ₂ -Z _n ) Z ₂₂ + (-z ₁₂ -ζ _π χζ ₂₁ + Z ₀ )} Z _rf3

+ {(-Z ₁₂ -Z,,) (Z ₂₁ + Z ₀ )-,! Z ₁₂ } Z ₂₂ -Z ,, Z ₁₂ (Z ₂₁ + Z ₀ )]

/ [{(Z ₂₂ + (Z ₂₁ + ZZ _ml3 + (-Z ₁₂ -Z _n ) Z ₂₂ + (-Z ₁₂ -Z „) (Z ₂₁ +2 ₀ )]

Fl f2 f3 is determined using at least one combination of Z, Z, and Z

The method of correcting a high-frequency characteristic error of an electronic component according to claim 1.

In the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample, the second correction data acquisition sample, and the electronic component are shunts. Connected,

In the above formula, the admittance Y when an electronic component is measured by the reference measurement system is measured.

m

And the admittance Y when the electronic component is measured by the actual measurement system.

m m d is derived on the basis of an error model connected to the determined terminal 1 2,

d d

When the admittance viewed from the terminal 1 is derived, the error model is such that the admittance Y is connected between the terminal 1 and the terminal 1, and the terminal 1 and the admittance Y

d Admittance Y is connected between the connection point of 12 m 12 and the ground, and the admittance Υ and the admittance 前記

11 12 Admittance Υ is connected between the connection point of terminal 1 and the ground, and the terminal 2 and the end d f d

An admittance Y is connected to the child 2, and the connection between the admittance Υ and the terminal 2 m 22 22 m

An admittance Y is connected between the point and ground,

twenty one

The admittance Υ Υ Υ Υ Υ

f 11 12 21 22

Y, Y, Y as a result of measuring the admittance of at least three of the first correction data acquisition samples in the first step; In the second step, the admittance of the terminal 1 is measured for at least three of the first correction data acquisition sample or the second correction data acquisition sample.

m

As a result of measuring the admittance of 端子, Υ, Y and the terminal 2 Υ, Y

mil ml 2 ml3 m m21

, Y and

ni2 m 23

The following formula [Equation 3a]

[Number 3a]

_{denom = (Y d2 -) Y} mU + (- Y d3) Y ml2 + (- Y d2) Y mU

^J ll = Soil Soil】 rfl V i】 Λ ^ mll sf

I denom

Yn

I denom The following formula woman 3b]

[ _Equation 3b] denom = (Y _dl ― Y _dl ) Y _m23 + (-Y _d3 ) Y _m22 + (one Y _d2 ) Y _m2l sj

-{(~

+ (-)

I aenom

Y ₂₂ = ± V ^ _2- (Ten (-Ten) (,) ((-

Of the 16 combinations of,, Υ, Υ, Y obtained from / denom,

11 12 21 22

For the following formula 4

[Equation 4] r _fl = ₂ + (Y ₂₁ + Γ ₀ )) ¾10 ((γ ₂₁ + γ ₀ ) + γ _η ) γ ₂ 2 + Άι + u

_{_{+ {(- 12 -γ η)}} γ 22 + (-Y -Y U) (Y 21 + Y 0)} Y DL

Tens {(— ¾ -Y) Y ₂₂ + (-Y _l2

+ {(--) + ₀ ) -7 „7 ₁₂ } 7 ₂₂ -Y _n Y _l2 (Y + Y ₀ )]

/ [{(¾ + (Y _2l + Y ₀ )} Y _m n + (--+ (-

Y _f , =-+ (Y _2l + Y ₀ )) + ((+) + + Yn (Y +)}

+ {(-Yn-) ¾ + (--W ₂₁

_{+ {(- -7 Η) (} 7 21 + Γ 0) -Κ "7 12} 7 22 -Y n Y 12 (Y 2, + Υ 0)]

_{/ [{(Υ 22 + (} Υ 21 +)} + (-¾ - 22 + (- - 21 +)]

Fl f2 f3 is determined using at least one combination of Υ, 致, and Υ

In the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample, the second correction data acquisition sample, and the electronic component are a series. Connected,

In the second step, the high-frequency characteristics prepared in the first step are different from each other, and at least three of the first correction data acquisition samples are priced in the reference measurement system, and in the second step For obtaining at least three second correction data that can be regarded as having at least three samples for obtaining the first correction data having different high-frequency characteristics or having the same high-frequency characteristics as the first correction data obtaining sample. A sub-step of converting a measured value of the sample in the actual measurement system into an impedance parameter and deriving a differential impedance component thereof;

The mathematical expression relates the impedance when an electronic component is measured by the actual measurement system to the impedance when the electronic component is measured by the reference measurement system, via a two-port error model. Yes, the impedance measured when measuring the electronic components with the reference measurement system is derived based on a one-port error model that has only one port to which two-port differential signals are input. The method of correcting a high-frequency characteristic error of an electronic component according to claim 1.

[5] In the actual measurement system, the first correction data acquisition sample and the electronic component, or the first correction data acquisition sample, the second correction data acquisition sample, and the electronic component are Shunt connected,

In the second step, the high-frequency characteristics prepared in the first step are different from each other, and at least three of the first correction data acquisition samples are priced in the reference measurement system, and in the second step For obtaining at least three second correction data that can be regarded as having at least three samples for obtaining the first correction data having different high-frequency characteristics or having the same high-frequency characteristics as the first correction data obtaining sample. A sub-step of converting the measured value of the sample in the actual measurement system into an admittance parameter and deriving its in-phase admittance component;

The mathematical expression relates the admittance when the electronic component is measured with two ports in the actual measurement system to the admittance when the electronic component is measured with the reference measurement system through the two-port error model. Yes, admittance is measured when electronic components are measured with the reference measurement system. The in-phase signal of two ports is derived based on a one-port error model that has only one port. The method for correcting a high-frequency characteristic error of an electronic component according to claim 1, wherein:

[6] The electronic component high-frequency characteristic error correction apparatus used in at least the fifth step of the electronic component high-frequency characteristic error correction method according to any one of claims 1 to 5, wherein the third step A storage unit for storing the mathematical formula determined in the above step and a measured value obtained by measuring the arbitrary electronic component obtained in the fourth step by the actual measurement system;

It can be obtained by performing calculation for correcting the measurement value stored in the storage unit using the mathematical formula stored in the storage unit and measuring the electronic component in the reference measurement system. An arithmetic unit for calculating an estimated value of the high frequency characteristics of the solder electronic component;

An electronic component high-frequency characteristic error correction device comprising: