JP4743208B2 - Method for measuring electrical characteristics of electronic components - Google Patents

Description

The present invention relates to a method for measuring electrical characteristics of an electronic component using a planar transmission line as a measuring jig, and more particularly to a method for measuring electrical characteristics such as impedance value and Q value of a single electronic component using a distributed constant type error model. Is.

As the operating frequency of high-frequency electronic circuits becomes higher and higher, electronic components (hereinafter referred to as devices) used in the circuit have to measure accurate electrical characteristics in the high-frequency region. Although it is generally difficult to measure high-frequency electrical characteristics of surface-mounted devices such as chip inductors and chip capacitors, measurement of high-frequency characteristics such as the impedance value and Q value of a single device is generally an impedance analyzer up to 3 GHz. Has been implemented.

In the measurement using an impedance analyzer, open correction and short correction are performed to determine the characteristics of a single device, and residual impedance and stray admittance of the test fixture and measurement cable are corrected. However, the impedance analyzer has a complicated apparatus, and it is difficult to cope with a high frequency, and it does not correspond to a frequency of 3 GHz or more. Therefore, a network analyzer is used for device measurement at a high frequency of 3 GHz or higher.

Device measurement using a network analyzer includes a one-port measurement method and a two-port measurement method. As a method for measuring a single device by 1-port measurement, a full 1-port calibration is performed at the tip of the coaxial cable, and then the device is measured via a jig that relays the coaxial cable and the surface-mounted device. As an example of such a measuring method, there is a method of obtaining characteristics of a single device by attaching an SMA connector, measuring the electrical length of the SMA connector, and performing open correction and short correction. However, accurate data cannot be obtained with such a measurement method. This is because, in the first place, the network analyzer is not good at measuring the reflection coefficient more than the transmission coefficient measurement, and it is very difficult to obtain the electrical characteristics of the SMA connector accurately. is there.

On the other hand, measurement with two ports can be measured with higher accuracy than with one port. As a device measurement method based on such 2-port measurement, TRL calibration and SOLT calibration are known as shown in Non-Patent Documents 1 and 2. In SOLT calibration, it is difficult to produce a standard device in the shape of a device on a planar transmission line at high frequencies exceeding 5 GHz, and high frequency open characteristics cannot be defined. There is a problem that it cannot be used for the above. Therefore, it is suitable to perform a high-frequency measurement of a surface mount device using a planar transmission line. There is a TRL calibration method or the like as a method for removing the error factor of the measurement system of the planar transmission path.

As described above, a network analyzer is used for device measurement at a frequency of 3 GHz or more. In measurement using a network analyzer, measurement using two ports is superior to measurement using one port. However, in a known calibration method such as the TRL calibration method, the calibration surface is calibrated, but no further calibration is performed. Of the calibration methods, the shunt method provides characteristics closer to the device characteristics than the series method. However, if the CPW width of the measurement board is different, the residual impedance differs, and this effect is reflected in the measurement value. Will also include the characteristics of the substrate.

Accordingly, the measurement value obtained by TRL calibration or the like includes device and substrate characteristics ahead of the calibration surface, and different data can be obtained depending on the type of calibration substrate. In general, in order to reduce the influence of the substrate characteristics as much as possible, a device such as making the substrate as thin as possible is devised. However, even if these devices are used, the obtained data is not the characteristics of the device alone.

In the case of 2-port measurement, the error cannot be expressed with a simple model as in the case of 1-port measurement. Therefore, correction using two parameters of residual impedance and floating admittance, open / short correction, It was not applicable.

Thus, since the high frequency data of a single device cannot be obtained, the following problem occurs.
(1) In the past, component manufacturers have used impedance analyzer data as the characteristics of a single device, but the high frequency data from a network analyzer of 3 GHz or higher includes device characteristics and board characteristics. There was a problem that data of a single device of 3 GHz or more could not be provided without tracing.
(2) In addition, when high-frequency data is used by a designer of an electronic device such as an electronic device set manufacturer as high-frequency data of a component, there is generally no problem if the design board used is the same as the high-frequency data. However, since the substrate to be used is usually different from the measurement substrate from which the high-frequency data is acquired, there is a problem that the design simulation cannot be performed with high-frequency data currently provided.
Agilent Technology Application Note 1287-9 “Fixture Measurement Using a Vector Network Analyzer” Agilent Technology Application Note 1463-5 “Evaluation of impedance characteristics of SMD parts using ENA and ICM test fixture”

Accordingly, an object of a preferred embodiment of the present invention is to provide a method for measuring the electrical characteristics of an electronic component, which can determine the electrical characteristics of a single subject with high accuracy.

The above object is achieved by a method for measuring electrical characteristics of an electronic component according to a preferred embodiment of the present invention.

In the first preferred embodiment of the present invention, the measurement error up to the calibration plane is corrected by an error correction method in which the subject is connected in series between the calibration planes of the plane transmission path and the plane transmission path is used as a calibration reference. A first step for determining the electrical characteristics of the subject, a second step for measuring the electrical characteristics of the planar transmission line in the open state, and a subject of the subject determined in the first step based on a distributed constant type error model A third step of correcting the electrical characteristics using the electrical characteristics in the open state measured in the second step, and a through chip pre-valued between the calibration surfaces of the planar transmission path are connected in series, and the planar transmission is performed. The fourth step of obtaining the electrical characteristics of the through chip in which the measurement error up to the calibration surface is corrected by the error correction method using the path as the calibration reference, and the distributed constant type error model A fifth step for correcting the electrical characteristics of the through chip obtained in the step 2 using the electrical characteristics in the open state measured in the second step, and a target obtained in the third step by using a distributed constant type error model. A sixth step of correcting the electrical characteristics of the specimen using the electrical characteristics of the through chip obtained in the fifth step and obtaining the electrical characteristics of the subject alone, and measuring the electrical characteristics of the electronic component Is the method.

In the second preferred embodiment of the present invention, the subject is shunt-connected to the calibration surface of the planar transmission path, and the measurement error up to the calibration plane is corrected by an error correction method using the planar transmission path as a calibration reference. A first step for obtaining the electrical characteristics of the specimen and a shunt connection of a short chip pre-valued to the calibration surface of the planar transmission path, and an error correction method using the planar transmission path as a calibration reference to the calibration plane. The second step for obtaining the electrical characteristics of the short chip in which the measurement error is corrected, and the short chip for which the electrical characteristics of the subject obtained in the first step are obtained in the second step by using a distributed constant type error model And a third step of obtaining the electrical characteristics of a single subject by correcting the electrical characteristics using the electrical characteristics of the electronic component.

In high-frequency measurement using a planar transmission path using two ports of a network analyzer, it is possible to correct a measurement error up to the calibration plane, but it is not possible to correct a measurement error between calibration planes. Therefore, measurement values obtained by TRL calibration and the like include not only the characteristics of a single object but also substrate characteristics between calibration surfaces, and different data can be obtained depending on the type of calibration board. Therefore, in the present invention, the measurement error between the calibration surfaces is removed from the measured value of the subject by performing distributed constant type open correction and through correction. In the 2-port measurement, it is possible to identify an unknown model parameter from the measured values in the open state and the through state by using the fact that four S-parameter measured values are obtained in one measurement, and mathematically remove this. It was to be. In this way, the electrical characteristics of a single subject that is not affected by the substrate characteristics can be obtained. In the method of the present invention, the error correction method for correcting the measurement error up to the calibration surface is not limited to a known method such as the TRL calibration method, and any correction method may be used as long as it is a two-port series method. Can do.

For open correction, the electrical characteristic S _OPEN may be measured using a network analyzer in a state where nothing is connected between the tips (calibration surfaces) of the flat transmission line. In the distributed constant type error model, the characteristic S _OPEN in the open state can be considered to be connected in parallel to the subject when the subject is measured. In order to remove this error factor from the measured object value, first, the S parameter (scattering coefficient matrix) is converted into the Y parameter (admittance coefficient matrix), and the open measured value is subtracted from the measured object value. Errors can be removed. Thereafter, if the Y parameter is inversely converted to the S parameter, the distributed constant type open correction is completed.

On the other hand, in the through correction, the electrical characteristic S _THRU may be measured using a network analyzer in a state where through chips that have been pre-valued are connected in series between the tips (calibration surfaces) of the flat transmission line. As the through chip, a chip having the same shape as the subject and capable of being connected to the same position as the connection position of the subject is preferable. In the distributed constant type error model, the characteristic S _THRU in the through state can be considered to be connected in series to the subject when the subject is measured. The characteristic S _THRU in the through state is also obtained from the electrical characteristics of the through chip with the measurement error up to the calibration surface corrected by the error correction method using the flat transmission line as the calibration reference, as in the case of the characteristic measurement of the subject. The electrical characteristics of the through chip may be obtained by open correction. In order to remove the error factor in the through state from the measured object value, the S parameter is converted into the T parameter (transmission coefficient matrix), and the measured object value is divided by the through measured value. it can. Thereafter, if the T parameter is inversely converted to the S parameter, the distributed constant type through correction is completed, and the scattering coefficient _SD of the subject alone can be obtained.

Even if the impedance Z _D is obtained by converting the scattering coefficient S _{D of the} object into the Z parameter, this value is a value that ignores the characteristic impedance Z _O of the planar transmission line and ignores the impedance of the through chip. . Therefore, when these impedances cannot be ignored, the impedance Z _DUT of the subject alone can be obtained by multiplying Z _{D by} the characteristic impedance Z _O of the planar transmission line and adding the impedance of the through chip. .

Although the above description is a correction method in a planar transmission line used in the series method, the present invention can also be applied to a planar transmission line used in the shunt method. In other words, when measuring by connecting a subject to a flat transmission line used in the shunt method, the error correction due to the capacitive component generated between the subject and the flat transmission line cannot be removed by the conventional correction method. It is not possible to accurately determine the characteristics of Therefore, in order to remove this error factor, the electrical characteristic S _SHORT in a state where the short chip is shunt-connected to the calibration surface of the planar transmission path is measured. In the case of a planar transmission line used in the shunt method, since the signal conductor is conducted between the ports, measurement in the open state is unnecessary. A short chip having the same shape as the subject and capable of being connected to the same position as the measurement position of the subject is preferable. The error factor that occurs between the short chip and the planar transmission line when the short chip is shunt-connected can be regarded as almost the same as the error factor that occurs between the subject and the planar transmission line when the subject is shunt-connected. . Therefore, the measurement error is identified from the electrical characteristic S _SHORT in a state where the short chip is shunt-connected, and the measurement error is removed from the measured value of the subject. In this way, the electrical characteristics of a single subject that is not affected by the substrate characteristics can be obtained. In the present invention, in the 2-port shunt method, it is possible to obtain the electrical characteristics of a single object with the influence of the substrate characteristics alleviated.

In the correction by the shunt method, first, the measurement data (corrected to the calibration surface) S _{A of the} subject and the measurement data (corrected to the calibration surface) S _SHORT with the short chip connected are obtained, and these are obtained as Z parameters. To Z _A and Z _SHORT . By adding the impedance of the short chip to the difference Z _D , the impedance Z _DUT of the subject alone can be obtained. If the short chip impedance is negligibly small, the operation of adding the short chip impedance to Z _D can be omitted.

Effects of preferred embodiments of the invention

As described above, according to one embodiment of the present invention, in the flat transmission line used in the series method, the data S _OPEN that measures the characteristics of the open state of the flat transmission line using a distributed constant type error model; Corrections are made using data S _THRU that measures the characteristics of a through-chip connected in series to a flat transmission line, so that measurement errors between calibration surfaces can be corrected, and the electrical characteristics of a single subject can be accurately measured. Can be sought.

According to another embodiment of the present invention, in a planar transmission line used in the shunt method, an electrical characteristic S _SHORT in a state where a short chip is shunt-connected to the calibration surface of the planar transmission line is measured, and a distributed constant type error is measured. Since the correction is performed using the electrical characteristic S _{SHORT according} to the model, the measurement error on the calibration surface can be corrected, and the electrical characteristic of the single object can be obtained with high accuracy.

Embodiments of the present invention will be described below with reference to examples.

The measurement method of this embodiment is an example of a 2-port series method using a network analyzer, and is referred to herein as the RRRR method.

−Preparation of planar transmission line−
The measurement jig 10 is the same as the measurement jig used in the TRL method, and as shown in FIG. 1, the two signal conductors 12a and 12b are arranged on a straight line on the upper surface of the dielectric substrate 11, and one end is spaced. A CPW (Coplanar Waveguide), which is a planar transmission line in which the ground conductors 13a and 13b are arranged at intervals on both sides in the width direction of the signal conductors 12a and 12b, is used. A ground conductor may also be provided on the back surface of the substrate 11. Connectors 14 and 15 are attached to both ends of the measuring jig 10 in the length direction. The signal lines 14a and 15a of the connectors 14 and 15 are connected to the signal conductors 12a and 12b, and the GND portions 14b and 15b are connected to the ground conductors 13a and 13b. Are connected to each. Connectors 14 and 15 are connected to measurement ports 18a and 18b of network analyzer 18 via coaxial cables 16 and 17, respectively.

-Connection and measurement of short circuit reference-
In this measurement method, the calibration standards to be measured are all the same short-circuit standard 20, and the measurement jig 10 to be used is the same jig. The short-circuit standard 20 refers to a general component in an electrical short-circuit state, and may be a chip component, a metal piece, a tool, or the like. Desirably, the contact length in the longitudinal direction of the transmission path such as a knife edge is short. If the short-circuit standard is ideal, the reflection coefficient will be -1 (total reflection). However, even though the short-circuit standard actually has some inductance, the inductance value needs to be known. It is. Usually, in the microwave band, the short-circuit state can be relatively easily obtained as compared with the open state. When high measurement accuracy is required, the short-circuited reference inductance may be obtained by simple simulation or the like.

When the transmission line characteristic of the measuring jig 10 is known, the error coefficient is obtained by short-circuiting the short-circuit reference 20 at three locations of the transmission line. When the transmission line characteristic is unknown, the short-circuit reference 20 is used as the transmission line characteristic. By short-circuiting at four locations, the transmission line characteristics and the error coefficient can be obtained simultaneously. Here, in order to simplify the explanation, a case where the transmission path characteristic is known will be described as an example. However, a derivation method when the transmission path characteristic is unknown is Japanese Patent Application No. 2005-44916, which is a prior application of the present applicant. Please refer to.

First, the signal conductor 12a and the ground conductor 13a or 13b are short-circuited by the short-circuit reference 20 at a location (measurement point P1 in FIG. 1) where the subject is connected during measurement of the subject, and this point is used as a calibration surface. In this example, the short-circuit reference 20 short-circuits the signal conductor 12a and both ground conductors 13a and 13b. However, the signal conductor 12a and one ground conductor may be short-circuited. The measurement result at this time is S _11M1 and the true value of the reflection coefficient at the measurement point 1 is Γ _A1 . Γ _A1 is the true value of the short-circuit reference, but this may be −1 if the length of the short-circuit reference 20 in the length direction of the transmission line is sufficiently small compared to the measurement signal wavelength, and otherwise. In addition, the expected value of the short-circuit reference inductance should be obtained by simulation or the like.

Next, the measurement is connected between the short circuit reference 20 and the signal conductor 12a grounding conductor 13 or 13b at the position spaced by a transmission line L ₁ from the measurement point P1 to the port 1 side (measurement point P2), when the The measurement result is S _11M2 . At this time, the true value of the reflection coefficient of the short-circuit reference 20 at the measurement point P2 is Γ _A1 , but when the measurement point P1 is taken as the reference plane, the true value of the reflection coefficient is converted as shown in Equation (1). This is because the electromagnetic wave incident from the port 1 side is totally reflected by the short-circuit reference 20, and therefore, the distance for transmitting the transmission path by 2L ₁ for the round trip is shorter than when the short-circuit reference 20 is connected to the measurement point P1. Here, α is the transmission rate [U / mm] per unit length, β is the phase constant [rad / mm] of the transmission channel, and α and β are known. Γ _A2 is the true value of the short-circuit reference 20 connected to the measurement point P2 when the measurement point P1 is taken as the reference plane.

_Subsequently , measurement is performed with the short-circuit reference 20 connected to a position (measurement point P3) on the transmission line that is L ₂ away from the measurement point P1 on the port 1 side, and the measurement result at this time is S _11M3 . When the measurement point P1 is taken as the reference plane as in the case of the measurement point P2, the true value Γ _A3 of the reflection coefficient is expressed by the following equation (2).

-Through-chip connection and measurement-
Next, as shown in FIG. 2, measurement is performed in a through state (direct connection between ports). An appropriate device (hereinafter referred to as a through chip) 21 for connecting the ports is connected in series between the signal conductors 12a and 12b. Measurements, reflection coefficient S _11MT, in S _22MT, the transfer coefficient and S _21MT, S _12MT. It should be noted that the electrical characteristics of the through chip 21 in the through measurement may be unknown, for example, a chip resistance whose resistance value is not known, but the transmission coefficient should not be directional. This condition is usually satisfied automatically because the transfer coefficient does not have directionality according to the reciprocity theorem unless a special material such as ferrite under a DC magnetic field is used.

-Calculation of error coefficient of error model for RRRR calibration-
An error model of RRRR calibration is shown in FIG. This is not particularly novel, and is the same as the error model for TRL correction that has been conventionally used. In the figure, S _11M and S _21M are measured values of the reflection coefficient and the transmission coefficient, and S _11A , S _12A , S _21A and S _22A are the true values of the scattering coefficient of the object. Further, although there are eight error coefficients E _xx and F _xx , the scattering coefficient measurement is a ratio measurement, and therefore, seven error factors may be determined. Specifically, E ₂₁ = 1 may be set.

Now, it is necessary to obtain each error coefficient in FIG. 3 from the measurement result obtained by connecting the short-circuit reference 20 described above. First, E ₁₁ , E ₂₁ , E ₁₂ , E ₂₂ , F ₁₁ , F ₂₁ , F ₁₂ , F ₂₂ is obtained by the following equation. Incidentally, F _xx are same as E _xx, describes only the E _xx. At this stage, for E ₂₁ and E ₁₂ , the product of both is obtained, but these cannot be obtained independently. D ₁ is an intermediate variable.

Next, the measurement results S _21MT and S _12MT of the forward and reverse transfer coefficients of the through chip can be written as follows using the error factors of FIG. However, the true value of the scattering coefficient of the through chip 21 is assumed to be S _11A , S _21A , S _12A , S _22A .

Here, the ratio of S _21MT and S _12MT is considered. Based on the formula (4), the following formula can be obtained by arranging the transmission coefficients in the forward and reverse directions of the through chip 21 with attention being equal (S _21A = S _12A ). It should be noted here that the scattering coefficients S _11A , S _21A , S _12A , and S _22A of the through chip 21 are all eliminated by division. In other words, even if the true value of the scattering coefficient of the through chip is unknown, if the through chip has no directionality, if the ratio of S _21MT and S _12MT (this is a measurable amount) is known, the relationship of the error coefficient is It is to be decided.

Based on Equation (3) and Equation (5), the total error coefficient can be determined as follows:

With the above, all error coefficients have been determined. The above is a discussion when a signal is applied from the port 1 side to the port 2 side (forward direction), but the reverse direction can be derived by setting F ₂₁ = 1 instead of E ₂₁ = 1.

-Measurement of subject and implementation of RRRR calibration-
When the error coefficient is obtained, the subject 22 is connected to the measurement jig 10 and its characteristics are measured. For example, the subject 22 is adsorbed using a chip mounter or the like, and the subject 22 is brought into contact with the subject measurement position of the measurement jig 10 to measure the electrical characteristics (S _11M, S _21M, S _12M, S _22M ). To do. At this time, when the subject 22 has two terminals, the signal conductors 12a and 12b may be connected in series as shown in FIG. 4A. However, when the subject 22 has three terminals or four terminals, FIG. What is necessary is just to connect between signal conductor 12a, 12b and grounding conductor 13a, 13b like (b). Therefore, the RRRR measurement method can be applied not only to a two-terminal electronic component but also to an electronic component having three or more terminals such as a filter.

Since the error model for RRRR calibration is the same as the error model for TRL correction, the same calculation as TRL correction may be performed to remove the effect of error from the actual measurement result of the object, and the effect of error is removed. The mathematical formula is described below. This equation is calculated based on the reflection coefficient in the case of 2-port measurement, but may be calculated from the four receiver outputs of the network analyzer in order to eliminate the influence of error factors. Also, in the case of three or more ports, the same equation as this equation may be used, or the influence of the error factor may be removed using a circuit simulation method. In short, any known technique may be selected. In Equation (7), D ₂ is an intermediate variable.

−Distributed constant open correction−
In the above-described RRRR calibration, the error factor up to the subject measurement position (calibration surface P1) can be removed, but the error between the subject measurement positions, that is, the error factor between the calibration surfaces is not considered. That is, as shown in FIG. 5, residual impedance, floating admittance, etc. existing between the calibration surfaces of the measurement jig 10 which is CPW are not corrected. Although FIG. 5 shows the error factor of the capacitive lumped constant, the open correction described below can be applied to any distributed constant circuit.

The error model of the distributed constant type open correction is such that the open characteristic is connected in parallel to the subject when the subject is measured. In the open measurement, a transmission line characteristic in a state where nothing is connected to the measurement jig 10 is measured, RRRR calibration is performed, and error corrected characteristics S _OPEN (S _11OPEN , S _21OPEN , S _12OPEN , S _22OPEN ) are obtained. This is an open characteristic between calibration surfaces.

The calculation of the open correction can be easily performed by converting the S parameter (scattering coefficient) into the Y parameter (admittance coefficient). In order to convert an arbitrary scattering coefficient matrix S into an admittance coefficient matrix Y, the following equation may be used.

The scattering coefficient matrix S _A of the subject is converted into an admittance coefficient matrix Y _A , and the scattering coefficient matrix S _OPEN of the open measurement result is converted into an admittance coefficient matrix Y _OPEN by Expression (8). Incidentally, the scattering coefficient matrix S _A of the subject is to connect the object 22 as previously described transmission line, which was obtained by performing the RRRR calibration, by the characteristics of the subject to remove error factor up to the calibration surface is there. Then, an admittance coefficient matrix Y _{D from} which the influence of open characteristics is removed is obtained by the following equation.

以上のようにして、オープン特性の影響を除去した被検体の特性Ｓ_B を求めることができる。The open correction is completed when the admittance matrix Y _D of the subject obtained by the above equation is converted into the scattering coefficient matrix S _B by the following equation.

As described above, it is possible to determine the characteristics of the subject to remove the influence of the open characteristics S _B.

−Distributed constant type through correction−
With the distributed constant type open correction, it was possible to obtain the characteristics of the subject from which the influence of the open characteristics was removed. However, it is considered that the characteristics thus obtained are still different from the characteristics of the subject alone. This is because there is an error factor that cannot be corrected by open correction, but occurs when a short-circuit between calibration surfaces is made. That is, when the subject is connected between the calibration surfaces, the capacity of the corner of the transmission path when the transmission path width and the subject width are different, the ground plane residual inductance between the calibration planes, and the like can be considered. Such characteristics occur in a distributed constant manner.

Therefore, a distributed constant type through correction is introduced. The error model of the distributed constant type through correction is a state in which the subject characteristic and the through characteristic are connected in series when the subject is measured. Usually, since mounting lands for two-terminal electronic components are generally designed to be symmetrical with respect to each port, it is assumed that error factors are symmetrical including reflection characteristics.

As shown in FIG. 6, the measurement is performed with the calibration surfaces connected by the through chip 23, and RRRR calibration is performed to obtain the characteristics after error correction. This data is further _subjected to the above-described distributed constant type open correction to obtain characteristics S _THRU (S _11THRU , S _21THRU , S _12THRU , S _22THRU ). This characteristic S _THRU is a through characteristic after the measurement error up to the calibration plane is corrected (for example, RRRR calibration) and the open correction is performed between the calibration planes. Here, the through chip 23 is priced in advance, and preferably has the same shape as the subject and can be connected to the same position as the connection position of the subject. In practice, it can also be used as the through chip 21 implemented in the RRRR calibration.

The error model of the through correction, since considered the object property S _B at the subject measurement and through characteristic S _THRU are connected in series, it is easy to calculate by converting the S parameters to T-parameters. For conversion from the S parameter to the T parameter, the following equation may be used.

The equation (11), the transmission coefficient matrix T _B the object property S _B, converts respectively the through characteristic S _THRU transmission coefficient matrix T _THRU. Here, object property S _B is measured by connecting the object 22 between the calibration surface, perform RRRR calibration, and a characteristic after the distributed constant type open correction, calculated by the equation (10) Equal to the characteristic S _B.

A transmission coefficient matrix T _{DUT from} which the influence of the through characteristic is removed is obtained by the following equation.

以上のようにして、オープン特性およびスルー特性の影響を除去した被検体単体の特性Ｓ_DUT を求めることができる。Finally, the T parameter is converted into the S parameter by Equation (13) to obtain S _DUT .

As described above, it is possible to obtain the characteristic S _DUT of a single object from which the influence of the open characteristic and the through characteristic is removed.

By the way, the measured value S _THRU in the through state is a value when L = 0 (H) of the through chip 23 connected between the calibration surfaces. Therefore, the characteristic S _DUT obtained by measuring the subject is smaller than the true value of the subject by the L value of the through chip 23. Therefore, in order to correct the true value of the subject, the L value of the through chip 23 is obtained in advance by an electromagnetic simulator, and the L value is corrected at the stage of _{obtaining the} final subject impedance Z _DUT as will be described later. Just do it.

The true values S _11A and S _21A of the reflection coefficient obtained by removing the error factor by RRRR calibration, open correction, and through correction can be converted into the Z parameters Z _11A and Z _21A of the subject by the following equations.

ここで、Ｚ₀ は特性インピーダンス（Ω）、ｆは周波数（Ｈｚ）、Ｌはスルーチップのインダクタンス値（Ｈ）である。被検体の対称性によりＺ_11AとＺ_21Aは同じ値になるため、Ｚ_DUT を求めるためには、数式（１５）のどちらを用いても構わない。なお、Ｓ_11A 、Ｓ_21A 、Ｚ_11A 、Ｚ_21A およびＺ_DUT はいずれも複素数である。The impedance Z _DUT of the subject can be obtained by multiplying the Z parameters Z _11A and Z _21A and the characteristic impedance Z _{0 of the} substrate by values at the same frequency. At this time, the true value of the impedance Z _DUT can be obtained by correcting with the inductance of the through chip 23.

Here, Z ₀ is a characteristic impedance (Ω), f is a frequency (Hz), and L is an inductance value (H) of the through chip. Since Z _11A and Z _21A have the same value due to the symmetry of the subject, either of the equations (15) may be used to obtain Z _DUT . S _11A , S _21A , Z _11A , Z _21A and Z _DUT are all complex numbers.

In the above calculation formula, it is considered that the through chip 23 has the characteristics as an inductor. However, when more strict accuracy is required, a more accurate equivalent circuit model (L and It is needless to say that it is desirable to obtain the impedance of the R series circuit and the like.

In the first embodiment, the error correction method has been described by taking the RRRR calibration series method as an example. However, the error correction method may be a known method such as the TRL calibration series method, distributed constant type open correction, distributed constant type. Through the through correction procedure, the characteristics of a single subject can be extracted with higher accuracy than before.

The measurement method of this embodiment is an example of a 2-port shunt method using a network analyzer, and is referred to herein as a TRRR method.

−Preparation of planar transmission line−
FIG. 7 shows a state in which the measurement jig 30 is connected to the network analyzer 38. The measuring jig 30 is formed by forming a planar transmission line composed of one signal conductor 32 and two ground conductors 33a and 33b on the upper surface of a dielectric substrate 31. In this example, a CPW (coplanar waveguide) in which the signal conductor 32 is continuously formed in the length direction of the dielectric substrate 31 and the ground conductors 33a and 33b are formed at intervals on both sides in the width direction of the signal conductor 32. It was used. Note that a ground conductor may also be provided on the back surface of the measurement substrate 30. Connectors 34 and 35 are attached to both ends of the measuring jig 30 in the length direction. Signal lines 34a and 35a of the connectors 34 and 35 are connected to the signal conductor 12, and GND portions 34b and 35b are connected to the ground conductors 33a and 33b. It is connected. The connectors 34 and 35 are connected to the measurement ports 38a and 38b of the network analyzer 38 via coaxial cables 36 and 37.

-Connection and measurement of short circuit reference-
Similarly to the above-described RRRR method, in this measurement method, when the transmission line characteristics of the measuring jig 30 are known, the error coefficient is obtained by short-circuiting the short-circuit reference 40 at three locations on the transmission line. The short-circuit standard 40 may be the same as the short-circuit standard used in the RRRR method. If the transmission line characteristic is unknown, the transmission line characteristic and the error coefficient can be obtained at the same time by short-circuiting the short-circuit reference 40 at four locations on the transmission line. Here, in order to simplify the explanation, a case where the transmission path characteristic is known will be described as an example. However, a derivation method when the transmission path characteristic is unknown is Japanese Patent Application No. 2005-44916, which is a prior application of the present applicant. Please refer to.

First, the signal conductor 32 and the ground conductors 33a and 33b are short-circuited by a short-circuit reference 40 at a location (measurement point P1 in FIG. 7) where the subject is connected during measurement of the subject, and this point is used as a calibration surface. The measurement result at this time is S _11M1, and the true value of the reflection coefficient at the measurement point P1 is Γ _A1 . Next, the short-circuit reference 40 is connected between the signal conductor 32 and the ground conductors 33a and 33b at positions (measurement points P2 and P3) on the transmission line that are separated by L ₁ and L ₂ on the port 1 side from the measurement point P1. Measurement is performed, and the measurement results at this time are _designated as S _11M2 and S _11M3 . The true values Γ _A2 and Γ _A3 of the reflection coefficient of the short-circuit reference 40 are obtained in the same manner as in the equations (1) and (2).

-Measurement in the through state-
Separately from the measurement based on the short-circuit reference 40, measurement is performed in a through state (direct connection state between ports). In the through state, the measurement is actually performed without connecting anything to the measurement jig 30. Measurements, reflection coefficient at S _11MT, the transfer coefficient and S _21MT.

-Calculation of error coefficient of TRRR calibration error model-
An error model of TRRR calibration is shown in FIG. This error model is the same as the error model of FIG. In the figure, S _11M and S _21M are measured values of the reflection coefficient and the transmission coefficient, and S _11A , S _12A , S _21A and S _22A are the true values of the scattering coefficient of the object. Further, although there are eight error coefficients Exx and Fxx, since the scattering coefficient measurement is a ratio measurement, it is only necessary to determine seven error factors. Specifically, E ₂₁ = 1 may be set.

Now, it is necessary to obtain each error coefficient in the figure from the measurement result obtained by connecting the short-circuit reference 40 described above. First, E ₁₁ , E ₁₂ , E ₂₂ , F ₁₁ , F ₂₁ , F ₁₂ , and F ₂₂ are mathematical expressions. It is obtained in the same manner as (3). In TRRR calibration, since the ideal through-state scattering coefficients S _11MT and S _21MT can be measured, F ₂₂ and F ₁₂ can be obtained by the following equations.

As described above, all error coefficients could be determined. The above is a discussion when a signal is applied from the port 1 side to the port 2 side (forward direction), but the reverse direction can be derived by setting F ₂₁ = 1 instead of E ₂₁ = 1.

-Measurement of subject and implementation of TRRR calibration-
When the error coefficient is obtained, the subject 41 is shunt-connected to the signal conductor 32 of the measurement jig 30 and one of the ground conductors 33a or 33b as shown in FIG. 9, and its characteristics are measured. For example, the subject 41 is adsorbed by using a chip mounter or the like, and the subject 41 is brought into contact with the subject measurement position (P1) of the transmission path 30 to obtain electrical characteristics (S _11M, S _21M, S _12M, S _22M ). Can be measured. At this time, the measurement jig 30 to be used is the same as that used in the TRRR calibration, and the measurement jig 30 and the coaxial cables 36 and 37 are also connected.

Since the error model for TRRR calibration is the same as the error model for TRL correction, the same calculation as that for TRL correction may be performed in order to remove the influence of the error from the actual measurement result of the object. Although the calculation formula for removing the influence of the error is the same as Formula 7, the calculation formula is not limited to Formula 7, and any known technique may be used.

−Distributed constant type short correction−
In the above-described TRRR calibration, the error factor up to the subject measurement position (calibration surface P1) can be removed. However, the error factor due to the capacitance component generated between the subject 41 and the planar transmission path on the calibration surface P1, ie, FIG. As shown, the stray capacitance C component generated between the electrode of the subject 41 and the pattern of the measurement jig 30 cannot be removed. As shown in FIG. 10, this error factor is removed by shunting an appropriate device (hereinafter referred to as a short chip) 42 to the calibration surface P1. The short chip 42 is priced in advance, and preferably has the same shape as the subject 41 and can be connected to the same position as the measurement position of the subject 41. More preferably, an electronic component of the same type as the subject 41 is preferable. That is, the C component generated when the subject 41 is connected is considered to be the same as the C component generated when the short chip 42 is connected. Measurement is performed with the short chip 42 shunt connected to the calibration surface P1, and TRRR calibration is performed to obtain error corrected characteristics S _SHORT (S _11SHORT , S _21SHORT , S _12SHORT , S _22SHORT ).

の計算を行い、被検体４１を接続した時に発生する誤差要因を除去したＺ_D を求める。Next, the value S _A obtained by shunt-connecting the subject 41 to the calibration surface and TRRR calibrated and the value S _SHORT obtained by shunt-connecting the short chip 42 to the calibration surface and converted to TRRR are converted into Z parameters, respectively. Z _A and Z _SHORT

To calculate Z _{D from} which an error factor generated when the subject 41 is connected is removed.

By the way, the measured value S _SHORT in the short state is a value when the L value of the short chip 42 connected to the calibration surface is assumed to be zero. Therefore, the characteristic S _A obtained by measuring the subject 41 is smaller than the true value of the subject 41 by the L value of the short chip 42. Therefore, in order to correct the true value of the subject 41, the L value of the short chip 42 is obtained in advance by an electromagnetic field simulator, and the L value is obtained at the stage of _{obtaining the} final impedance Z _DUT of the subject 41 as will be described later. Correction may be performed.

ここで、Ｚ_O は測定治具の特性インピーダンス（Ω）、ｆは周波数（Ｈｚ）、ＬはショートチップのＬ値（Ｈ）である。被検体の対称性によりＺ₁₁とＺ₂₁は同じ値になるため、Ｚ_DUT を求めるためには、数式１８のどちらを用いても構わない。As shown in the following equation, Z _D reflection coefficient and transmission coefficient Z ₁₁ , Z ₂₁ obtained by removing the error factor by Equation 17 and the characteristic impedance Z _{0 of the} substrate are multiplied by values at the same frequency, respectively. At the same time, by correcting with the L value of the short chip 42, it is possible to obtain the impedance Z _DUT value of the subject in which the influence of the measurement substrate and the short chip is mitigated.

Here, Z _O is the characteristic impedance (Ω) of the measuring jig, f is the frequency (Hz), and L is the L value (H) of the short chip. Since Z ₁₁ and Z ₂₁ have the same value due to the symmetry of the subject, either of Equation 18 can be used to obtain Z _DUT .

In the above calculation formula, it is considered that the short chip 42 has a characteristic as an inductor. However, when more strict accuracy is required, a more accurate equivalent circuit model (L and It goes without saying that an R series circuit or the like can be used.

As described above, in the 2-port series measurement method, by performing distributed constant type open correction and distributed constant type through correction, residual impedance and floating admittance between calibration surfaces can be corrected, and the characteristics of a single object can be extracted. I was able to do that. In the 2-port measurement of the network analyzer, it is possible to obtain the high-frequency characteristics of a single object that does not include board characteristics. Therefore, 3 GHz that is traced with the measured value of the impedance analyzer that has been used as a characteristic of a single object by a component manufacturer. It has become possible to provide high-frequency data from the network analyzer to users.

Further, in the two-port shunt measurement method, by performing distributed constant type short correction, it is possible to remove an error factor that occurs between the subject and the planar transmission line in a state where the subject is shunt-connected. In this way, the electrical characteristics of a single subject that is not affected by the substrate characteristics can be obtained.

However, in order to trace the measurement value obtained by the impedance analyzer and the measurement value obtained by the network analyzer according to the present invention with high accuracy, the measurement state of the measurement jig, for example, a mechanism for holding the subject or a positioning mechanism, is the same. The conditions such as the definition of the equivalent circuit model of the through chip used in the through correction being the same are necessary.

In addition, when a designer of an electronic device such as an electronic device set manufacturer uses high-frequency data of a single subject according to the present invention as a design simulation, a technique for superimposing necessary circuit board parameters and the characteristics of the subject according to the present invention. By using, it becomes possible to reproduce a highly accurate simulation.

It is a top view of an example of the measuring apparatus in the 2 port series method concerning the present invention. It is a top view of the measuring apparatus in the through measurement concerning this invention. It is an error model figure used with the RRRR calibration method concerning this invention. It is a top view at the time of the subject measurement of the measuring device concerning the present invention. It is a top view which shows the influence of the error factor of the open state which generate | occur | produces between calibration surfaces. It is a top view which shows the influence of the error factor of the through state which generate | occur | produces between calibration surfaces. It is a top view of an example of the measuring device in the 2 port shunt method concerning the present invention. It is an error model figure of TRRR calibration concerning the present invention. It is an enlarged view of the state which connected the subject to the measurement jig. It is an enlarged view of the state which connected the short chip to the measuring jig.

Explanation of symbols

10, 30 Measuring jig (planar transmission path)
12a, 12b Signal conductors 13a, 13b Ground conductors 18, 38 Network analyzer 20, 40 Short circuit reference 21 Through chip 22, 41 Subject 23 Through chip 32 Signal conductors 33a, 33b Ground conductor 42 Short chip

Claims (7)

A first step of obtaining electrical characteristics of a subject in which a measurement error up to the calibration plane is corrected by an error correction method using a series connection of the subject between calibration planes of the plane transmission path and using the plane transmission path as a calibration reference. When,
A second step of measuring the open state electrical characteristics of the planar transmission line;
A third step of correcting the electrical characteristics of the subject obtained in the first step by using the electrical characteristics in the open state measured in the second step by a distributed constant type error model;
Through chip electrical characteristics obtained by correcting the measurement error up to the calibration plane using an error correction method using the plane transmission path as a calibration reference by connecting a series of through-chips pre-valued between the calibration planes of the plane transmission path. A fourth step for determining
A fifth step of correcting the electrical characteristics of the through chip obtained in the fourth step by using the electrical characteristics in the open state measured in the second step by using a distributed constant type error model;
According to a distributed constant type error model, the electrical characteristics of the subject obtained in the third step are corrected using the electrical characteristics of the through chip obtained in the fifth step, and the electrical characteristics of the subject alone are obtained. And a method for measuring electrical characteristics of the electronic component. In the third step, the measurement data S _A of the subject obtained in the first step and the data S _OPEN in the open state obtained in the second step are converted into Y parameters, and Y _A , Y _OPEN age,
Y _B = Y _A -Y _OPEN
Seeking Y _B performs calculation,
The method of measuring an electrical property of an electronic component according to claim 1, wherein the Y _B is inversely converted into an S parameter to obtain a scattering coefficient S _B of the subject after open correction. In the fifth step, the through-chip measurement data S _C obtained in the fourth step and the open state data S _OPEN obtained in the second step are converted into Y parameters, and Y _C , Y _OPEN age,
Y _THRU = Y _C -Y _OPEN
Y _THRU is calculated by calculating
3. The method for measuring electrical characteristics of an electronic component according to claim 2, wherein the Y _THRU is inversely converted into an S parameter to obtain a through-chip scattering coefficient S _THRU after open correction. T _B and T _THRU are obtained by converting the scattering coefficient S _B of the subject after the open correction and the scattering coefficient S _THRU of the through chip after the open correction into T parameters,
T _D = (√T _THRU ) ⁻¹ T _B (√T _THRU ) ⁻¹
Or
T _D = (√T _B ) ⁻¹ T _THRU (√T _B ) ⁻¹
I asked to T _D carried out of the calculation,
The T _D is inversely converted into S-parameter, electrical characteristic measuring method of the electronic component according to claim 3, wherein the determination of the scattering coefficient S _D of the object after the open corrected and slew correction. By converting the scattering coefficient S _D of the subject into Z parameters, multiplying the converted Z _{D by} the characteristic impedance Z _O of the planar transmission line, and adding the impedance of the through chip, the impedance of the subject alone _{5. The} method for measuring electrical characteristics of an electronic component according to claim 4, wherein Z _DUT is obtained. A first step of obtaining an electrical characteristic of the subject in which the measurement error up to the calibration plane is corrected by an error correction method using the plane transmission path as a calibration reference by shunt-connecting the subject to the calibration plane of the planar transmission path; ,
Shunt connection of a short chip that has been pre-valued to the calibration surface of the planar transmission path, and the electrical characteristics of the short chip that has corrected the measurement error up to the calibration plane by using the planar transmission path as a calibration standard. A second step to seek;
Third, the electrical characteristics of the subject obtained in the first step are corrected using the electrical characteristics of the short chip obtained in the second step, and the electrical characteristics of the subject alone are obtained using the distributed constant type error model. And a method for measuring electrical characteristics of an electronic component comprising the steps of: The electrical property S _A of the subject obtained in the first step and the electrical property S _SHORT in the short state obtained in the second step are converted into Z parameters to be Z _A and Z _SHORT ,
Z _D = Z _A −Z _SHORT
7. The electrical characteristic measurement of the electronic component according to claim 6, wherein Z _D is obtained by performing calculation and the impedance Z _DUT of the subject is obtained by adding the impedance of the short chip to Z _D. Method.

Images