Data processing: measuring – calibrating – or testing – Measurement system – Performance or efficiency evaluation
Reexamination Certificate
2000-05-22
2002-11-19
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system
Performance or efficiency evaluation
C324S076190
Reexamination Certificate
active
06484124
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to an improved system for measuring selected performance characteristics of electronic components. In one preferred embodiment, the present invention comprises a method and apparatus for evaluating selected performance criteria of microwave power components, and in particular, microwave transmitter and receiver components. The present invention comprises improvements that individually or in combination increase the accuracy, reproducibility, and speed of these determinations, in a system that is lightweight, portable, and low cost, relative to prior known devices.
BACKGROUND OF THE INVENTION
Microwave components have long been critical features of radar systems, electronic devices, and other systems. Errors in the parameters of microwave components translate directly into decreased accuracy and precision of the equipment, systems, and processes in which they are employed. There has long been a need to improve the accuracy, reliability, repeatability, and correlation of signal-to-noise ratio (“SNR”) measurements of microwave power transmitter and receiver components. Prior to the present invention, a relatively high degree of variability existed between test sites, as well as between test sets at the same test site. Improvement in the accuracy of the performance characteristics of microwave components contributes directly to improved accuracy and precision in the systems in which they are used. Prior approaches have not adequately met this need. In response to this need, a self-calibrating measurement technique having improved speed, accuracy, repeatability, and portability of the present invention has been developed.
Many microwave transmitters have stringent requirements on their output signal-to-noise ratio (“SNR”). For example, in a pulsed radar system, the transmitter SNR can limit the ability of the radar to detect and/or track weak targets against a background of echoes from the earth's surface (clutter). The techniques commonly used prior to the present invention to measure transmitter and microwave tube SNR can have errors of ±1.5 dB. As a result, there has been poor agreement among SNR measurements made at the component factory, transmitter manufacturer, system integrator, and end user. This lack of agreement causes disputes over product acceptance, results in rejection of components meeting specification and acceptance of components failing to meet specification, creates unnecessary product returns, and generates wasteful requests to retest components. Moreover, the prior known methods for determining the performance characteristics of microwave power components, and in particular SNR, were: cumbersome; time consuming; expensive; and involved the use of large, heavy, bulky, and expensive equipment to carry out the analysis.
Prior to the present invention, microwave components have been tested in a test setup of the type depicted in the right-hand portion of FIG.
1
. The right-hand portion of
FIG. 1
depicts a test set up used for evaluating a component under test, in this example, a microwave power tube (or transmitter). A radio frequency (rf) drive signal was supplied to the component. The rf output from the component under test was applied to a vector demodulator circuit which, converted it into in-phase (I) and quadrature (Q) video signals. The rf drive signal to the component was also used as the reference (local oscillator) input to the demodulator as shown in FIG.
1
. The signals were then digitalized and filtered. In the test setup depicted in
FIG. 1
, a Tektronix RTD-710A digitizer, sampling at 50 Msa/sec, was used to convert the I and Q video signals into 10-bit digital format. Analog 5 MHZ 6-pole Bessel filters at the inputs of the digitizer channels were used to define the bandwidth in which the tube noise is measured, i.e. ±5 MHZ about the carrier. For pulsed radar tubes, the digitized data was captured in a 2-microsecond window located near the middle of the pulse. The raw data was then transferred to a general purpose digital computer. Application software compensated the data for pulse-to-pulse variations caused by the modulator, and computed the average intrapulse SNR referenced to a 1 MHZ bandwidth. This conventional technique, however, took substantial amounts of time per measurement and resulted in SNR measurements with an error of ±1.5 dB.
The inventor observed that accuracy of prior known methods was limited by vector demodulator errors, which could be removed by lengthy calibration procedures. Specifically, there are both linear and non-linear error sources in the measurement of microwave power components.
Linear Distortion: An ideal vector demodulator (VDM) generates a unit circle centered at the origin on the IQ plane. The I and Q video output voltages are defined by the equations:
I=kA
cos &thgr;
Q=kA
sin &thgr;
where
A=rf signal voltage
k=mixer conversion loss
&thgr;=phase angle of rf signal with respect to the LO
Plotting I and Q data from a real vector demodulator generated an elliptical locus displaced from the origin as shown in FIG.
2
. The inventor observed that this linear distortion resulted from three error sources:
DC Offset—Imbalances in the mixer diodes and transformers create low level dc outputs in the I and Q channels when the local oscillator signal is applied. This causes displacement of the center of the locus from the origin. These offsets are a function of the measurement frequency and, if not compensated, cause a few tenths of a dB error.
I/Q Channel Gain Imbalance—This difference in gain between I and Q channels changes the locus into an ellipse with its principal axes parallel to the I and Q axes. Factors which contribute to gain imbalance include variations in:
rf power split into the I and Q channel mixers;
VSWR of the mixer rf ports;
mixer conversion loss;
insertion loss of video filters;
digitizer channel gains;
The first three items were considered by the present inventor to be dominant, and varied as a function of the test frequency.
Quadrature Error—Quadrature error causes the principal axes of the elliptical locus to tilt relative to the I/Q axes and results from the I and Q channel phases differing from 90 degrees. Differential phase errors are caused by:
The rf power splitters used to supply RF and LO to the I and Q mixers;
Rf line lengths; and
Mixer VSWRs.
These errors are also functions of the test frequency. These three linear distortions were determined by the inventor to be the major causes of SNR measurement inaccuracy.
The distorted I and Q voltages take the form:
I
=
kA
(
1
-
C
2
)
cos
(
θ
-
D
2
)
+
B
cos
E
Q
=
kA
(
1
+
C
2
)
sin
(
θ
+
D
2
)
+
B
sin
E
where
C
=
gain imbalance;
D
=
quadrature error
B
=
dc offset amplitude
E
=
dc offset phase relative to the I axis
C
dB
=
2
log
(
1
+
(
C
2
)
1
-
(
C
2
)
)
FIG. 3
illustrates the linear distortion in a typical vector demodulator. A gain imbalance of ±1 dB and a phase error of ±6 degrees are typical values. These errors are attributed to mixer Voltage Standing Wave Ratios (“VSWRs”), which can be in the range 3.0:1. In addition, this vector demodulator (VDM) shows a bias in the phase of about −8 degrees. This is attributed to a small differential error (c. 0.05 inches) in the rf line lengths feeding the I/Q mixers. The dc offset for this VDM is on the order of 1%, and is a small error source.
Non-Linear Distortion—Prior to the present invention, the LO to signal ratio typically was not tightly controlled. The inventor observed that non-linear distortion generated in the demodulator could be made negligible by maintaining the signal amplitude 15 dB below the LO reference level. The LO/signal ratio was typically only 8 dB, where considerable amplitude compression was readily observable.
The calibration procedure of prior known methods typically
Collier Shannon Scott PLLC
Hoff Marc S.
Raymond Edward
Technology Service Corporation
LandOfFree
System for measurement of selected performance... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System for measurement of selected performance..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System for measurement of selected performance... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2961135