1 PCB board broadband parameter extraction
1.1 purpose of the experiment
This case mainly shows how to use ads to process the S parameters of the transmission line measured by the network analyzer, extract the plate medium parameters from the measured data, and improve the design accuracy. The experimental contents include writing AEL function, post-processing verification, and de embedding.
Through this case study, the engineer will be involved in the operation method
Using the AEL language to write equation for post-processing
How to use the de embedded module
Use and Simulation of S parameter file
1.2 background knowledge
1.2.1 de embedding technology
In the actual measurement process, the test results include the error caused by the test fixture. We can describe the measured part and fixture as a set of S parameters, as shown in Figure 1-1.
SD: represents the S parameter of the tested part
SA / sb: represents the S parameter of the fixture on the left and right sides of the tested part
Figure 1-1 measuring the measured part by fixture
De embedding technology is an operation process that separates the S parameters of the measured part from the total S parameters of the measured part and the fixture based on the known s parameters of the fixture (SA / sb).
The de embedding technique uses matrix calculation of circuit network. For the convenience of calculation, the S parameter is usually converted to the T parameter for cascade operation.
So for the device on the fixture, the final measurement result is
Therefore, as long as the T matrix of the test fixture is obtained in advance, its influence can be eliminated in the test
The General de embedding steps are as follows
A to obtain the mathematical model of the test fixture, the S parameter or T parameter can be used to represent each half of the test fixture
B do full dual port calibration and measurement for the vector network analyzer, and the measurement results include the data of the fixture and the measured part
C convert s parameter to t parameter
D use the de embedding formula to de embed the measurement results
E convert the result to S parameter SD.
1.2.2 AFR automatic fixture removal calibration technology
Automatic fixture removal (AFR) calibration technology is a simple method to extract accurate broadband fixture model. This calibration technique can be used in a variety of fixture and interconnection structures, such as adapter, chip package, cable, PCB printed transmission line and through hole. This calibration technology has the same high-precision calibration performance as the traditional TRL calibration technology, but it has a simpler fixture manufacturing implementation. AFR is widely used in the field of high-speed signal integrity.
Figure 1-2 test board
Figure 1-2 shows a test plate with a through hole as the tested part. Through hole structure as the object of study, in the middle of two uniform transmission lines, the two ends of the transmission line are SMA adapters, which are used to connect the network analyzer to measure the S parameters of the through hole. In this example, we are concerned with the through hole. For measurement, the through hole is in the middle of the fixture (the fixture includes SMA connector and transmission line connecting the through hole). From the blue TDR response curve in Fig. 1-2, it can be seen that SMA adapter brings discontinuity that cannot be ignored, and the transmission line is not completely uniform. In the process of transmission, the fluctuation of impedance can be observed, which also brings the transmission loss.
It is a common basic problem for all test and measurement. Just like the above example, how to separate the measurement result of the tested part from the whole measurement result (the tested part plus fixture). This is also the problem to be solved by AFR automatic fixture removal calibration technology.
At present, there are many calibration technologies for test measurement in the industry, including
Sol t (short circuit, open circuit, load, through)
TRL (through, reflection, transmission line)
LRM (transmission line, reflection, matching)
The above three calibration techniques make a compromise between the complexity of the implementation and the accuracy of the measurement results. Unfortunately, for TRL and LRM calibration techniques, which are widely used in test and measurement to remove the influence of fixture, there are high requirements for the correct fabrication of fixture in the implementation process, and it is easy to introduce artificial error. The higher the requirement of calibration accuracy, the more difficult it is to realize the fixture.
The automatic fixture removal (AFR) calibration technology has made a breakthrough in the relationship between the accuracy of calibration and the ease of implementation, which is the original contradiction. The design of the fixture is very simple while it has extremely high accuracy.
AFR technology is realized when the fixture on both sides of the tested part is mirror symmetrical. In this case, we need to make a fixture calibration part to extract the fixture s parameters. The form of the calibrator is to connect the clamps on both sides directly to form a straight through structure which is twice the length of the clamp on one side. This kind of calibrator is usually called 2x straight through reference fixture, as shown in Figure 1-3.
Figure 1-3 2x straight through reference fixture
Although the fixture on one side is not symmetrical, when two symmetrical fixtures are cascaded, the new 2x straight through reference fixture calibrator is mirror symmetrical. Therefore, in the S parameters of the calibrator obtained through the test, S11 = S22, S21 = S12, two known quantities can be obtained, which is not enough to solve the three unknown quantities of the S parameters of the fixture on one side (s21a = s12a), as shown in Figure 1-4.
Figure 1-4 s parameter cascade
AFR technology is based on the feature that the calibration part of 2x straight reference fixture contains a uniform transmission line in the middle, and s11a and s22b of fixture can be extracted by using time-domain signal processing method. With the help of an extra known quantity, the S parameter of one side fixture can be uniquely solved. Using the de embedding technology, the influence of fixture can be removed from the test results, and the S parameters of the tested parts can be obtained.
1.2.3 Svensson / Djordjevic medium model
The dielectric constant of a medium, also known as permittivity, is the ratio of potential shift d to electric field strength E. The loss of lossy medium is usually expressed by the loss tangent (tand) of imaginary part of complex permittivity.
In order to ensure the causality of media, sveson / Djordjevic model is used as the broadband model of lossy media in ads
Where FL and FH are the parameters of the model
When the frequency approaches infinity, the value of dielectric constant, a, is a constant. The two parameters, ER / tand / freqforepsrtand / lowfreqfortand / highfreqfortand, can be calculated from other parameters entered by users in ads.
Fig. 1-5 and Fig. 1-6 show the frequency dependence of the dielectric constant using the sveson / Djordjevic model
Er = 4.6
TanD = 0.03
FreqForEpsrTanD = 1 GHz
HighFreqForTanD = 1 THz
LowFreqForTanD = 1 kHz
Figure 1-6 the imaginary part of dielectric constant and the change of loss tangent with frequency
1.3 experimental steps
1.3.1 import and verification of measurement files
1 model preparation
The circuit used to extract the dielectric constant of PCB is a stripline with length L of 1 inch. In order to test the S parameter of the stripline, fixture a and fixture B composed of SMA adapter, pad and transmission line are respectively on both sides of the stripline. Fixture a and fixture B are designed as mirror symmetrical structure, as shown in Figure 1-7.
Use keysight’s network analyzer to test the S parameters of the circuit in Figure 1-7. The port impedance is 50ohm, and the measurement frequency range is 10MHz ~ 40GHz. Save the measurement results as a file in touchstone format_ 2p4in_ T.s2p。
Figure 1-7 circuit to be tested
In order to get the S parameter of the stripline with length L of 1 inch, it is necessary to remove the influence of the fixture on the S parameter. Here, the AFR technology introduced above is selected, and the S parameters of 2x straight through reference fixture connected with fixture a and fixture B need to be measured.
The designed 2x straight through reference fixture is shown in Figure 1-8. Fixture a and fixture B are mirror symmetrical. However, due to the instability and error in the manufacturing process of the circuit, the S parameters of fixture a and fixture B, as well as the S parameters of each fixture in Figure 1-7 and figure 1-8, are inconsistent to a certain extent.
Use keysight’s network analyzer to test the S parameters of the circuit in Figure 1-8. The port impedance is 50ohm and the measurement frequency range is 10MHz ~ 40GHz. Save the measurement results as a file in touchstone format_ 1p4in_ T.s2p。
Figure 1-8 2x straight through reference fixture
Open keysightads software and create a new project file_ Parm_ ExtracTIon_ Wrk, assuming that the directory of the project file is PCB_ Parm_ ExtracTIon_ wrk。
Put the. S2p file (as shown in Figure 1-9) in the compact disc file into the PCB_ Parm_ ExtracTIon_ In the wrk / data folder, there are four files.
Figure 1-9 measured S-parameter file
Verification of 2S parameters
New schematic A1_ meas_ Data, edit the schematic diagram according to figure 1-10, and edit the parameters of the following modules and simulators.
File=” MD_ 1p4in_ T.s2p”
File=” MD_ 1p4in_ T.s2p”
Figure 1-10 schematic diagram
Click the shortcut icon “simulate” to run the simulation. In the pop-up display window, use the shortcut icon “rectangular plot” on the left to display the S parameters of the 2x straight through reference fixture circuit and the tested part + fixture circuit (Fig. 1-11).
Figure 1-11 S-parameter simulation results
The S parameters in Fig. 1-11 are completely consistent with those of the network analyzer during the test. At the same time, it can be seen from the phase curve that the two groups of phases differ in the propagation delay caused by the tested part.
We can also verify the above conclusion from the time-domain curve and edit the equation fixture_ 2X_ TDR and fixture_ plus_ DUT_ TDR (Fig. 1-12), TDR curve in time domain is displayed by “rectangular plot” (Fig. 1-13), 2x straight through reference fixture has two obvious discontinuities (fixture a and fixture b), and the circuit delay of DUT is larger than that of straight through fixture.
Figure 1-12 formula
Figure 1-13 S-parameter TDR results
1.3.2 ARF and measurement results de embedding
1. Fixture s parameter extraction
Run the software in the main menu of “Automatic Test Wizard” (Fig. 1-14).
Figure 1-14 running AFR
In the pop-up window, set AFR at each step. On page 1, select single ended signal dual port measurement (Fig. 1-15).
Figure 1-15 AFR page 1
In page 2, select “2x thru” (Fig. 1-16), that is, fixture a and fixture B are straight through. In AFR technology’s double port straight through S parameter extraction, it is considered that S21 of fixture a and S12 transfer function of fixture B are the same.
Figure 1-16 AFR page 2
In AFR wizard page 3, load the S parameter file “MD” measured before_ 1p4in_ T.s2p”。 After load, the characteristic impedance of the transmission line in the fixture and the transmission delay of the fixture will be calculated at the bottom of the page, as shown in Figure 1-17.
Figure 1-17 AFR page 3
On page 4, select “de embedding for PLTs measurements” (Figure 1-18).
Figure 1-18 AFR page 4
On page 5, select according to figure 1-19, and select the output file name “my”_ If you click “save fixture files”, two. S2p files will be generated, representing the S parameters of the two fixtures.
Figure 1-19 AFR page 5