ECE 451 Automated Microwave Measurements Laboratory Experiment No.

9 (PART A) PORT EXTENSION METHOD AND TRL CALIBRATATION

Introduction In previous experiments, the one-port, 3-term vector error correction model was shown to be very effective in characterizing and remving systematic measurement errors. Wth this method, systematic errors are characterized through measurement of calibration standards, for which the actual responses are known. The accuracy obtainable is largely dependent upon the accuracy to which these actual responses are known. For coaxial measurements, suitable calibration standards with known im pedances are readily availableshort, open, and matched term inations are com monly used. In non-coaxial measurements, however, it is much more difficult to build standards with impedances which are easily characterized. This is particularly true in microstrip line, where short circuits are inductive, open circuits radiate energy, and purely resistive loads are hard to build. In this experim ent, you will measure the impedance of the microstrip student unknown shown in Figure 1. Like the coaxial student unknown (with which you are by now well familiar), this microstrip version consists of a reasonably well-matched termination and a

tuning stub, the latter providing a large impedance variation versus frequency. The stub is now open-ended, however, unlike the shorted coaxial stub. The load impedance is defined by the arbitrarily-chosen reference plane specified in Figure 1. Calibration of the m easurement system at th is ref erence plane will be performed by two methods in this experiment. For the first method, the familiar one-port calibration will be performed with coaxial standards, with electrical delay then added to effectively extend the measurement plane through the connector and out to the desired point on the m icrostrip line, thereby obviating the need for difficult-to-build microstrip standards. This has two potentially serious dif ficulties: (1) the coax-to-microstrip transition discontinuity, and (2) non-ideal transm ission-line characteristics of the microstrip line, such as attenuation and dispersion. Neither of these effects are characterized and removed by this so-called port extension calibration technique. The second method used in this experiment is the TRL (from "Thru-Reflect-Line") calibration method. The TRL method uses transmission line standards instead of discrete im pedance standards to obtain the terms of a two-port error m odel. For more information about this m ethod, refer to Hewlett-Packard Product Note 8510-8, Applying the HP 8510B TRL calibration for non-coaxial measurements. You will be given three m icrostrip standa rds for TRL calibration, THRU, REFLECT, and LINE. With the port definition of Figure 1, the THRU standard, at 3 in. long, is a zero-length connection between ports 1 and 2. The LINE standard (3.73 in. long) is designed to be one-quarter wavelength (i.e., 90 electrical degr ees) longer than the zero-length THRU at the center frequency of interest, here 3.0 GHz. The REFLECT standard, though not a perfect open, is specified as such; its impedance need only be known approximately. Procedure Part I (Port Extension Method) In this section, two steps will be perform ed. First, the electrical length of the LINE standard, to be used later with the TRL method, will be determined. Then, the microstrip student unknown impedance will be measured using the port extension method. 1) Press PRESET to set the VNA to a known state. 2) Perform an S11 1-PORT calibration for Port 1, covering the 1 - 5 GHz range (201 points). Choose Cal kit 1 or 2. 3) Verify that you have a good calibration by observing measured responses for all calibration standards. Make certain that any added ELECTRICAL DELAY is set to zero. Electrical Delay can be found in Scale > Electrical Delay 5) Now, the electrical length of the LINE standard, relative to the THRU standard, will be measured. To do this, perform the following steps: a) Connect the THRU standard to Port 1, terminating it with a male type SMA short. b) Set display to PHASE.

c) Under the TRACE menu, click on Math/Memory to store data into m emory. Press DATA -> MEMORY . In the trace window, display the trace from the Memory by selecting Memory trace in Trace view options (in the same Math/Memory dialog). d) Replace the THRU standard with the LINE standard, term inating again with a short. Open a New Trace (in the same window) by clicking Trace > New Trace >S11. e) Adjust ELECTRICAL DELAY so that the current DATA and MEMORY trace are identical (this should require between around 200 picoseconds of delay). f) Record the electrical delay for later use. It is equal to twice the electrical length of the LINE standard, relative to the THRU. 6) The m icrostrip student unknown impedance will be measured next. This will require extending the measurement plane from its current location to the reference plane of Figure 1. Perform the following steps: a) Connect the TRL REFLECT standard to Port 1. b) Set display to SMITH CHART. c) Dial in the appropriate amount of ELECTRICAL DELAY so that the displayed impedance is balled up around that of an open (somewhere between 500 and 550 picoseconds should work). d) Replace the REFLECT standard with the microstrip student unknown. e) Save the trace in your folder as *.s1p. Part II (TRL Calibration) In this section, the TRL m ethod will be used to m easure the im pedance of the m icrostrip student unknown. The results will be comPared with the port extension mEasurement performed above. Before using the TRL m ethod, the characteristics of the THRU, REFLECT, and LINE calibration standards m ust be entered into the VNA. Once this informAtion is entered, the VNA cal kit must be modified to specify the use of these newly-defined standards. This will all be done next. 1) To define the new calibration standards, first bring up the Calibration > Advanced Modify Cal Kit menu. A new pop-up window must show up. 2) Click Insert New, Give your best kit name like ECE_451_TRL. Say something on Kit description. 3) Add connectors, In the new pop-up window, fill SMA connector fam ily, Gender Male, frequency from 0 to 8999 MHZ. 4) At the bottom , click on Add and add OPEN and in the new pop-up window, fill the following in the appropriate place. Label REFLECT, set C0- C3 to 0. Delay is 0 psec. This will add REFLECT standard in to your cal kit. To add the THRU and LINE, choose THRU option. The only difference in between them is that you have to set the DELAY for the LINE to the value you got from PART I. Make sure you label the LINE as LINE (not as THRU). Now you m ust have three standards in your Edit kit mini window.

5) Now, in the same window change the Class Assignments to TRL then click edit, in the new pop-up window, assign each of the standards. For example, Choose TRL THRU in Calibration Kit Class and select and add THRU from the Unselected Standards to Selected Standards . Do the same process for TRL REFLECT and TRL LINE. In Calibration Reference Zo option, choose LINE Zo. In Test Port Reference Plane, choose THRU STANDARD. Then click Ok and exit from the Class Assignments window. Also click Ok and exit from Edit kit window. 6) Before you start calibration, set the Electrical Delay to zero. Now Perform Full TRL 2-PORT(1,2) calibration. Make sure to enable View or Select Cal kit . Choose your new modified cal kit. You can check your Class assignments are correct at this time. 7) Measure your Microstrip student unknown with port 1 and save the result in *.s1p. Now, you have two m easurement files of the same student unknown, one with port extension method and the other is from TRL calibration. Read the files with ADS and plot Log Magnitude, Phase and Smith Chart of the two measurements on the same axis. As mentioned above, the port extension calibration of Part I does not remove errors introduced by the coax-to-microstrip transition discontinuity, yet the Part I results agree fairly well with the TRL measurements. Questions:- Answer the following in your lab report. 1) When would we want to perform a TRL calibration instead of a SOLT calibration? 2) Must the standards used in a TRL calibration be more or less ideal than those used for SOLT calibrations? Explain. 3) What is the major assumption about the TRL standards? 4) Calculate the Electrical Length of your LINE standard assuming speed of light? 5) What is the theoretical optimal electrical length of LINE for the Start and Stop frequencies given in the procedure part? Do es this value agree with the calculated electrical length from question 4? What is the cause of the error, if any? 6) Calculate the phase degree at the Start a nd Stop frequencies. Does our LINE standard satisfy the recommended insertion phase requirements? 7) Be sure to explain any differences between the port extension and TRL calibration measurements observed on these plots. Specifically, what phenom enon observed on the port extension measurement is elim inated (or greatly reduced) with the TRL measurement? Why is it eliminated with TRL calibration? (Hint: Read the Agilent TRL Application Note in the class website)

(PART B) Calibration Error Investigation

In this experiment, the calibrations standards will be interchanged in various com binations to determine what effect is observed when measuring known loads. Be sure to record on the plot what standards were used and what load is being measured. Use 300 MHz - 1.3 GHz range. Calibrate 1-PORT SOLT and interchange standards as defined below: 1. Interchange a short with an open (50 ohm termination as usual). 2. Interchange a short with a 50 ohm termination (open as usual). 3. Interchange a 50 ohm termination with a 5.0 VSWR standard (short and an open as usual). For each of the above combinations, measure and record the Smith Chart data terminated with each of the following loads: 1. A short. 2. An open. 3. A 50 ohm termination. 4. Your student unknown.