Stable load pull operation using tuners

Abstract

Load Pull tuning pattern and probe movement algorithms allow creating a test pattern allowing to avoid instability regions and spurious oscillations of microwave transistors during testing using slide screw load and source tuners. The impedances are selected based on the stability circle and instability area on the Smith chart and the probe movement trajectory allows both avoiding the static and circumventing the transient crossing through the instability area. All tuning commands are saved in a pattern file.

Claims

1. A method allowing stable (spurious oscillation-free) load pull measurement of microwave transistors (DUT) using calibrated programmable electro-mechanical slide-screw impedance tuners, wherein the tuners comprise input (test) and output (idle) ports and a slotted airline between the ports; and at least one reflective probe remotely movable along and insertable into the airline slot, and appropriate probe positioning gear, control electronics, firmware and a system control computer; and wherein the tuner calibration data comprising s-parameters of the tuner for a multitude of probe positions are saved in a tuner calibration file; said method comprising a) generating and using reflection factor tuning pattern ensuring stable transistor DUT operation; b) tuner probe movement trajectory allowing spurious oscillation-free load pull; whereby the reflection factor pattern excludes points in the instability area of the transistor DUT, and wherein the movement trajectory of the probe(s) circumvents the instability area of the transistor DUT.

2. Reflection factor tuning pattern as in claim 1 comprising a multitude of reflection factor points, whereby each point included in instability area of the transistor DUT is eliminated from the pattern.

3. The tuner probe movement trajectory as in claim 2, whereby the probe is withdrawn from the slabline before moving horizontally.

4. The method as in claim 3, whereby the DUT is mounted in a test fixture of which the s-parameters of the input and output sections are saved in a fixture calibration file.

5. A method for creating the reflection factor pattern, as in claim 4, comprising the following steps: a) retrieve the tuner calibration data from file; b) retrieve s-parameter calibration data of the test fixture from file; c) cascade the s-parameters of the output section of the test fixture with the tuner calibration data and save in a setup calibration file; d) retrieve the s-parameter matrix of the DUT, [Sab] whereby {a, b}={1,2}; e) create a reflection factor () pattern comprising N points (.sub.1 . . . .sub.N); f) step, in computer memory, through all points .sub.i (1iN), calculate g) the input reflection factor .sub.in of the DUT .sub.in (Sab, .sub.i) and eliminate all reflection factors .sub.i, for which |.sub.in (Sab, .sub.i)|1; h) step, in computer memory, through all remaining points .sub.j (1jN) and calculate the associated horizontal and vertical coordinates (X.sub.j, Y.sub.j) of the tuner probe using the setup calibration data created in step c); i) add, after every set (X.sub.j,Y.sub.j), two sets of coordinates (X.sub.j,Y.sub.min), (X.sub.j+1,Y.sub.min), whereby Y.sub.min is the withdrawn vertical position of the probe, for which the reflection factor presented to the DUT at its output port is outside the instability area; j) save all sets of the probe coordinates of step i) as follows: (X.sub.1, Y.sub.1), (X.sub.1, Y.sub.min), (X.sub.2, Y.sub.min), (X.sub.2, Y.sub.2), (X.sub.2, Y.sub.min), (X.sub.3, Y.sub.min), (X.sub.3, Y.sub.3), etc., and the associated reflection factor in a pattern file.

6. The method as in claim 1, comprising the following steps: a) retrieve probe coordinates from the pattern file as in claim 5; b) move the probe to the coordinates retrieved in step a); c) measure RF and DC parameters (Gain, Power, Intermod, Efficiency, etc.) of DUT; d) save DUT parameters and associated reflection factors in a load pull data file.

Description

DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

(1) The invention and its mode of operation will be more clearly understood by the following description in view of the appended drawings, in which:

(2) FIG. 1 depicts prior art: the block diagram of a passive load pull test system using a source and a load impedance tuner. The tuners and instruments are remotely controlled by a computer.

(3) FIG. 2 depicts prior art: a schematic front view of a, single probe, motorized (automated) slide screw impedance tuner for microwave frequencies.

(4) FIG. 3 depicts prior art: a perspective view of a section through a slabline with center conductor and capacitive metallic probe (slug) and the key dimensions of the probe and possible movements and limitations.

(5) FIG. 4 depicts prior art: a cross section of the metallic probe (slug) approaching the center conductor in a slabline and a method to prevent mechanical contact of the probe to the center conductor using an electro-mechanical limit switch.

(6) FIG. 5 depicts prior art: the tuning range of a slide screw tuner at a given frequency in the Smith chart (reflection factor plan). The tuning range (shaded area) is reduced due to, among other, the insertion loss in connectors and transmission line (slabline). Prevalent relation is: Z=R+jX=Zo*(1+)/(1); Zo=50; =||*exp(j).

(7) FIG. 6 depicts prior art: a typical test setup used to characterize (calibrate) an RF slide screw tuner on a Vector Network Analyzer (VNA) by measuring its s-parameters as a function of probe coordinates (positions).

(8) FIG. 7 depicts Smith chart with indication of the stable and unstable DUT areas for a number of frequencies; the areas inside the circles are potentially unstable.

(9) FIG. 8 depicts tuning areas and patterns and tuning trajectory between points A and B, for avoiding crossing through unstable areas.

(10) FIG. 9 depicts a flowchart for generating the probe coordinates reflection factor trajectory and probe movement instructions in order to avoid crossing through high electric field areas.

(11) FIGS. 10(a) and 10(b) depict: 10(a) depicts definitions of tuner probe movement coordinates; 10(b) depicts the actual probe movement trajectory corresponding to FIG. 8.

DETAILED DESCRIPTION OF THE INVENTION

(12) The tuner must be pre-calibrated to be used. This is done using a setup as in FIG. 6: It comprises a pre-calibrated Vector Network Analyzer, VNA (100); a system control PC (102), which controls the tuner motors (109) via a digital connection (108), a motor driver electronic circuit (107) and digital cable (1002); RF cables (105) connecting the VNA (100) to the test (106) and idle (1003) tuner ports. The VNA is calibrated at the reference plane of the test port (106) and measures the reflection factor S.sub.11. The reflection factor .sub.in at the test port is equal to S.sub.11 when the impedance connected to the idle port is 50 (or .sub.load=0), see equation 1. The carriage (104) can travel horizontally and the probe (103) vertically approaching the center conductor of the slabline (1001). For each probe setting the s-parameters of the tuner are measured and saved in a calibration file for the given frequency. The probe settings are selected to cover the Smith chart. Since the test fixture (FIG. 1) introduces a phase shift, it is usually impossible to select probe states to cover a preset area of the Smith chart. Therefore, preferably, the whole chart is covered.

(13) There are a number of possibilities in generating, validating, saving and executing load pull measurements within the stable area of impedances. The term validating is used here to signify that the amplitude and phase of the reflection factors remain outside the potentially unstable area (171) in FIG. 7. This can be determined by calculating the reflection factor at the input of the DUT (.sub.in) as a function of the DUT s-parameters (S.sub.ij) and the .sub.load as follows:
.sub.in=S.sub.12*S.sub.21*.sub.load/(1S.sub.22*.sub.load){eq. 1}
The .sub.load points, for which |.sub.in|1, are potentially unstable. During the load pull operation .sub.load is created by the tuner S.sub.11 at its test port, as long as the tuner itself is terminated with 50; if this is not the case appropriate calibration of the setup components and corrections are applied. In a first embodiment the validated test impedances for a specific frequency and all associated motor control instructions for positioning the probe are saved in a script (MACRO) file as follows: The impedance points of the load pull impedance pattern are a) selected, b) validated concerning spurious oscillations, c) probe coordinates for all corresponding impedances are calculated, d) associated motor movement instructions, as described so far, are determined and e) saved in a MACRO load pull file (script file) to be retrieved and executed later in a load pull operation.

(14) In a second embodiment only the impedance points of the load pull pattern are saved in the MACRO impedance pattern file, for a specific frequency. This can be done by hand, point by point, or through a simple algorithm generating a grid of points within given boundaries and saved in the MACRO file. It is during the load pull operation that a) the impedance points of the MACRO impedance pattern file are retrieved, b) validated concerning risk for spurious oscillations, c) the associated probe positions are calculated, using the tuner and test fixture calibration data, and d) the load pull operation (probe movement and data acquisition and saving) is executed.

(15) As already discussed before the probe movement shall not cross an area in the Smith chart, that may cause transient oscillations. The concept is shown in FIG. 8. It is in general ignored that a slide screw tuner, in order to tune between two points A and B, which are both inside the stable region of the Smith chart, may have to travel on an unstable trajectory (182) covering more than 180 degrees; this happens if the two points, unbeknown to the user, lie one at the beginning phase of the calibration (here point A) and the other at the end (here point B). If this happens, a transient oscillation may occur, during the movement of the tuner carriage/probe during the crossing through the instability area. This must be avoided. The only way is to create a safe trajectory.

(16) The solution disclosed here consists of creating a probe movement pattern file which will allow the tuner to move from point A (with probe coordinates (X.sub.A,Y.sub.A)) to point B (with probe coordinates (X.sub.B,Y.sub.B)) without crossing through the instability area. This is the trajectory (183): From the point A, in a first step, the probe is withdrawn, (first half of the trajectory (183), between points A and C) reducing this way the reflection factor within a safe area (181). This corresponds to a partially or totally withdrawal of the probe at a penetration depth Y=Ymin: the tuner moves from point A (X.sub.A,Y.sub.A) to point C (X.sub.C,Y.sub.min). Both points are stable. Then the carriage moves horizontally on the (stable) circle (181) from X.sub.C to X.sub.B until it reaches the expected horizontal position (X.sub.B) corresponding to point B. Then the probe is inserted to the final depth Y.sub.B following the second part of the trajectory (183), between points C and B.

(17) The method for creating the reflection factor pattern required for stable operation comprises, therefore, the following steps: a) The tuner calibration data are retrieved from stored file; b) S-parameter calibration data of the test fixture are retrieved from stored file; c) The s-parameters of the output section of the test fixture are cascaded with the tuner calibration data and saved in a new setup calibration file; d) S-parameters of the DUT are retrieved; e) A reflection factor () pattern is created that contains up to N points (.sub.1 . . . .sub.N); typical values for N is between 50 and 100 points; f) A software routine steps through all points .sub.i(1iN) and calculates the reflection factor .sub.in at the DUT input as a function of i, and eliminates all .sub.i for which |.sub.in(.sub.1)|1; the remaining points are renumbered then to .sub.j, whereby 1jM and whereby MN; g) Another software routine steps through all points .sub.3 (1jM) and calculates the associated horizontal and vertical coordinates (X.sub.j,Y.sub.j) of the tuner probe using the setup calibration data created in step c); h) Subsequently, two sets of coordinates (X.sub.j,Y.sub.min), (X.sub.j+1,Y.sub.min), are added after every remaining (X.sub.j,Y.sub.j) set to the total list of coordinates, whereby Y.sub.min is the vertical position of the probe, for which the reflection factor presented to the DUT at its output port is small (approximately 0.1 or less); i) All sets of the probe coordinates of step h) are then saved as follows: (X.sub.1,Y.sub.1), (X.sub.1,Y.sub.min), (X.sub.2,Y.sub.min), (X.sub.2,Y.sub.2), (X.sub.2,Y.sub.min), (X.sub.3,Y.sub.min), (X.sub.3,Y.sub.3), etc, together with the associated reflection factor , in a pattern file; that means, each time before the carriage moves horizontally, the probe is withdrawn to reduce the reflection factor presented to the DUT during the horizontal movement. The actual pattern generation algorithm is shown in the flowchart of FIG. 9; the final probe movement is shown in FIG. 10(b) whereby FIG. 10(a) clarifies the definition of the probe movement coordinates; every time the probe dips to the actual position a measurement is taken and the probe is withdrawn before the carriage moves horizontally.

(18) The section of the test fixture inserted between the tuner test ports and the DUT terminals (FIG. 1) determines the actual reflection factors, both presented to the DUT and calculated at the input tuner test port. It needs to be clarified that, for the DUT to actually oscillate both ports must be terminated within the unstable region. Since, however all components in the setup are wideband, i.e. the DUT has considerable gain at any frequency, from DC to high GHz range, as shown by the wideband (frequencies F1, F2, F3, F4 . . . ) stability circles in FIG. 7, that the tuners are typically also wideband, covering several frequency octaves, the test procedure must proceed with caution. It is not secure enough to determine that a specific set of source and load reflection factors satisfies the stability conditions at a specific frequency, since it is possible that at another frequency this will not be the case. This is why the proposed solution errs towards safety. The test fixture s-parameters (FIG. 1) are embedded into the tuner calibration files using s-parameter cascading and then these new data are used as expanded tuner calibration data. Alternatively the test fixture s-parameters can be cascaded with the DUT s-parameters and used together with the original tuner calibration data. The stability conditions are either valid or invalid independently on the used reference plane.

(19) The parameters to be measured by the load pull system vary depending on the application. The method disclosed here is valid for measuring any DC or RF parameter; their specific definitions are irrelevant to the method and they will always be falsified by the DUT spurious oscillations.

(20) Obvious alternative algorithms of the above method and dynamic probe movement limiting mechanisms are imaginable, but their effect will always be to introduce stable test conditions for all probe trajectories. The obvious variations, however, shall not impede on the validity and general coverage of the claims of the hereby described concept of using a dynamic moving trajectory routine for preventing the probes in slide screw tuners from creating transistor instabilities.