Electricity: measuring and testing – Impedance – admittance or other quantities representative of... – Parameter related to the reproduction or fidelity of a...
Reexamination Certificate
2001-01-16
2003-05-27
Metjahic, Safet (Department: 2858)
Electricity: measuring and testing
Impedance, admittance or other quantities representative of...
Parameter related to the reproduction or fidelity of a...
Reexamination Certificate
active
06570394
ABSTRACT:
BACKGROUND—FIELD OF THE INVENTION
This invention relates to systems for testing signal paths. More particularly, it relates to systems for measuring a non-linear distortion created by a reference test signal on signal paths. More particularly, it relates to systems for measuring a signal level at which an upstream cable television system begins to distort the signals that are being transported over it because a composite signal level is too high.
BACKGROUND—DESCRIPTION OF PRIOR ART
When setting signal levels for a transmission system it is important to operate with a high transmit power to have a best possible signal-to-noise ratio at a receiver. However, operating above the linear range of the transmission system creates high levels of non-linear distortion, which interfere with clear signal reception. The presence of linear distortion complicates the measurement of non-linear distortion. Whereas a percentage of non-linear distortion is dependent on input signal level, a percentage of linear distortion is not. Thus, there is a need to test non-linear distortion on a variety of level-sensitive signal paths including upstream cable television systems, high-powered high-definition television transmitters, microwave signal links, satellite transponders, and audio sound systems, so that the transmit signal levels can be set accurately.
A signal path may be represented as having a gain or a loss and both linear and non-linear distortion components. Amplifiers or filters with non-flat amplitude or non-linear phase response may create linear distortion components. Linear distortion components may also be created by multi-path distortion, which is also known as echoes or ghosts. Non-linear distortion components may be created by devices such as amplifiers with limited dynamic range, and laser diodes that extinguish if overdriven, or loudspeakers which have magnets that saturate if the loudspeaker is overdriven.
Non-linear distortion may be represented by a Taylor series that represents a transfer function of output voltage as a function of input voltage:
v
out
(
t
)=
Av
in
(
t
)+
Bv
in
(
t
)
2
+Cv
in
(
t
)
3
+ . . . (1)
Where v
out
(t) is the output of the signal path, v
in
(t) is the input signal, A is the fundamental coefficient which is amplification, B is the second order distortion coefficient, and C is the third order distortion coefficient. The coefficients with a higher order than three may or may not be significant, depending on the nature of the signal path. If balanced push-pull amplifiers create the non-linear distortion, the even order distortion components, such as the second order distortion coefficient, B, tend to be greatly reduced.
The problem of finding the non-linear distortion characteristics of a signal path is greatly influenced by whether or not the signal path is wide bandwidth or narrow bandwidth. Wide bandwidth systems, such as cable TV systems or audio systems, pass more than an octave of bandwidth. If clipping occurs in a narrow bandwidth system, such as a high-powered broadcast transmitter, the distortion components will be close in frequency to the carrier. Typically the third-order distortion component will be the dominant component in narrow bandwidth systems.
Determining the dynamic range of a signal path is not a new problem. One solution that has been employed in the past on wide bandwidth systems is to transmit a continuous wave (CW) test signal (a sinewave) into a signal path, such as a cable television upstream system. The transmitted continuous wave is adjusted upward and downward in level at a house while observing the output of the signal path on a spectrum analyzer located in a headend. A headend is the origination point for downstream cable television signals and a termination point for upstream signals. The test signal level at which non-linear distortion is created may be noted as the level at which harmonics of the fundamental continuous wave are created, or cross some predetermined threshold. This test method requires that the bandwidth of the signal path be wide enough to pass harmonics of the fundamental test signal. For example, if the bandwidth of a cable system upstream signal path is 5 to 42 MHz, a CW signal of 8 MHz can be applied to the input of the signal path. If the signal path is being clipped, a second harmonic at 16 MHz, a third harmonic at 24 MHz, a fourth harmonic at 32 MHz, and a fifth harmonic at 40 MHz will all be created according to the Taylor series expansion of equation (1).
In a narrow bandwidth system, the above technique will not work because the bandwidth is too narrow to pass the higher-order harmonics.
The continuous wave can be used in a narrow bandwidth signal path to test the clipping point of a signal path by plotting the magnitude of the output signal level on an abscissa of a graph (Y-axis) vs. the magnitude input signal level on an ordinate of a graph (X-axis). The input and output levels may be measured on a spectrum analyzer. When the output level is low, there is little distortion, and an X-Y plot will be a straight line. As the input signal level is elevated there will be a point where the output signal level can no longer increase at the same rate as the input signal level because of clipping or non-linear distortion. An output signal level that is 1 dB lower than the straight line is a commonly used reference level, and is called the 1 dB compression point. Since magnitude measurements are used, the trace will be located in one quadrant of a Cartesian coordinate system.
Another similar prior art method that is used to test narrow bandwidth signal paths is to apply two CW test signals that are close in frequency to the signal path. As the two CW signals are elevated in level, distortion components at a pair of sum and difference frequencies will be created. If one continuous wave is at a frequency f
a
and the other is at a frequency f
b
, the distortion components will be at frequencies of 2f
a
−f
b
and 2f
b
−f
a
. These distortion products are typically generated by the third order distortion coefficient, C, in the Taylor series expansion.
When both ends of the wide bandwidth signal path are at the same physical location, the clip-point of a wide bandwidth signal path can be observed by exciting the signal path with a sine wave signal. A high-speed oscilloscope is put into an X-Y mode where the input signal is displayed on the X-axis and the output signal is displayed on the Y-axis. The signal path is excited by a CW signal. If there is no non-linearity, an X-Y trace will be a straight line. If there is non-linearity, the transfer function can be observed on the oscilloscope as a bending of the trace. One disadvantage of this technique for measuring non-linear distortion is that all linear distortion, including delay, must first be removed. Another disadvantage of this technique is that signals from both ends of the signal path under test, which may be a network with delay, must be connected to the oscilloscope. Yet another disadvantage of this technique is that the signal path's bandwidth must be wide bandwidth (wide enough to pass a fundamental component of the continuous wave signal, plus several harmonics that were created by non-linear distortion acting upon the fundamental component). The advantage of this method is that the oscilloscope trace shows a X-Y plot that displays the transfer function, illustrating the non-linear distortion. The transfer function can be evaluated to determine the coefficients (A, B, C, etc.) of the Taylor series expansion. The X-Y plot will have magnitude and phase information on both the input and output signals and will not be constrained to one quadrant of the Cartesian coordinate system.
Another prior art method that is used to test the dynamic range of a signal path is to use a broadband source of random, or Gaussian, noise followed by a notch filter device. This technique works for both wide bandwidth and narrow bandwidth systems. The noise source, which has a portion of its energy removed in a narrow fr
Greene Kevin E.
Kerveros J
Lacasse Randy W.
Lacasse & Associates LLC
Metjahic Safet
LandOfFree
Tests for non-linear distortion using digital signal processing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Tests for non-linear distortion using digital signal processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Tests for non-linear distortion using digital signal processing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3020695