Electricity: measuring and testing – Measuring – testing – or sensing electricity – per se – Frequency of cyclic current or voltage
Reexamination Certificate
1999-12-15
2004-04-20
Metjahic, Safet (Department: 2858)
Electricity: measuring and testing
Measuring, testing, or sensing electricity, per se
Frequency of cyclic current or voltage
C324S076190
Reexamination Certificate
active
06724178
ABSTRACT:
BACKGROUND OF THE INVENTION AND THE PRIOR ART
Distortion performance of RF components subjected to broadband multichannel signal inputs are often measured by using a Multicarrier Generator (“MCG”) as a signal source. The measurement practice typically involves feeding the MCG's composite signal to a Device Under Test (“DUT”) and observing its output signal with a spectrum analyzer in a way that permits the observation and measurement of additional spectral components that are generated due to nonlinear distortions of the DUT. Of particular importance are measurements of broadband active devices' second and third order distortion components. These are called the Composite Second Order (“CSO”) and Composite Triple Beat (“CTB”) distortion components.
Prior art practices for measuring these distortion components using non-coherent MCG are described in detail in measurement standards adopted by the Society of Cable Telecommunications Engineers (“SCTE”) and are available as documents entitled “
Composite Triple Beat Distortion
”, IPS-TP-206, SCTE (Oct. 31, 1997) and “
Composite Second Order Distortion
”, IPS-TP-207, SCTE (Oct. 31, 1997). These practices are designed to provide with distortion measurement methods that can closely predict actual performance of active devices in cable TV systems.
Most cable systems and some MCGs that emulate cable systems are non-coherent systems in which individual carrier frequencies are not rigidly related to each other and may each independently vary over a frequency range of hundreds or thousands of Hertz relative to their nominal frequency setting. When the carriers are unmodulated in such non-coherent systems, specific distortion components (CTB or CSO on any particular channel) constitute narrow-band signals that may each consist of hundreds or even thousands of distortion signal terms spread out in frequency over several kHz. This necessitates the setting of the Spectrum Analyzers' Resolution Bandwidth (“RBW”) to 30 kHz and performing video filtering with a low video bandwidth (10 Hz or 30 Hz), and video averaging if possible.
It is important to note that both video filtering and video averaging applied in such measurements essentially amount to time-averaging of the output of the spectrum analyzer's LOG amplifier which is fed by its IF envelope detector. Hence, the practice in the industry is to report the average of decibel values of the fluctuating distortion power rather than its average power in decibels. It can be shown mathematically that absent such time averaging (i.e. video bandwidths settings that exceed the RBW), the first order probability density function of such measured results is a Log-Rayleigh distribution and that the variance is approximately 5.6 dB, independent of the absolute levels, the channel or even the order of the distortion term.
The results under video filtering conditions depend on many factors including the spectral distribution of the distortion signals. In this context, if the distortion power spectra does indeed fall well within the 30 kHz RBW, and at the same time has a smooth spectral characteristics devoid of pronounced power variations over a frequency scale of less than the video filter bandwidth, then one can obtain a reasonably accurate and stable measurement of the average distortion power.
Typically however, the fine structure of the spectral distribution of distortion terms is unknown and may vary from one instance to another which may result in loss of both the accuracy and repeatability of the measurement. If distortion terms fall outside of the analyzer RBW setting, the analyzer will consistently underestimate the true distortion power. Alternatively, if a significant portion of the distortion power spectrum has pronounced spectral power variations over a frequency range smaller than the video filter bandwidth, then distortion measurements will not be repeatable as a consequence of insufficient video averaging of very slow fluctuations. Ironically, this phenomena of slow fluctuation in the averaged distortion power is more pronounced with improved frequency precision of the non-coherent carriers, as the distortion components are dispersed over a narrower bandwidth, giving rise to large spectral power variations over a narrower frequency range.
The slow fluctuation and lack of repeatability of these distortion measurements was recognized and prior art methods attempting to mitigate it have been reported in a conference paper entitled “
CTB/CSO Measurement Repeatability Improvements Using Uniformly Distributed Noncoherent Carrier Frequencies
”, by E. J. McQuillen and D. Schick, published in the Proceedings of the SCTE Emerging Technologies Conference, pp 315-328; San Antonio, Jan. 28-30, (1998). These authors proposed a “Pseudorandom Spreading” method of intentionally dispersing the actual frequencies of all the carriers by pseudorandom frequency deviations of up to a few kHz so that the resulting distortion components would appear spread out over a frequency range that is up to three times wider than that, thereby reducing the likelihood of slow distortion envelope fluctuations.
One of the difficulties with such a “Pseudorandom Spreading” method is that by its very nature, it spreads out the distortion spectra away from the center of the Resolution Bandwidth Filter. The 30 kHz RBW filter mode used in the spectrum analyzer has a 3 dB bandwidth of 30 kHz, which means that a frequency response loss of 1-2 dB can easily be incurred for these dispersed distortion components. This factor can cause a systematic error by underestimating the distortion power. Indeed, the above referenced paper's authors themselves report without any explanation a measurement bias of 2 dB as compared to the non-dispersed case. Furthermore, the actual bias depends on the specific tone that is being measured and the specific collection of terms and their respective frequency deviations from the center of the filter. Alternatively, Expanding the RBW might reduce this bias but it will be at the expense of noise immunity.
In other approaches, prior art use of coherent sources for distortion tests was also made but for the reasons discussed below was often met with significant inconsistencies and deviations from expected results. One type of a coherent MCG source differs from non-coherent head-ends and simulators in that it generates an Incrementally Related Coherent (“IRC”) multicarrier signal. The multicarrier signal is generated in accordance with an IRC frequency plan in which carrier frequencies f
n
are given by the following formula:
f
n
=n
·6 MHz+1.2625 MHz,
where n represents the carrier index. Thus, carriers are spaced by 6 MHz and fall at offsets of 1.2625 MHz relative to 6 MHz multiples. For test purposes, an MCG in which n takes on values between 9 and 135 is preferable. All carriers generated by such coherent source are locked to a common signal reference. Small deviations in the reference frequency will result in small deviations in the carrier spacing and offset. However, these deviations will be scaled for all channels with the same scale factor. Thus, all channels will still be spaced by exactly a common frequency spacing and will be located at the same fixed frequency offset relative to multiples of the carrier frequency spacing. The coherent MCG can be based, for example, on the apparatus which can generate a plurality of IRC signals with very low phase noise as described in U.S. Pat. No. 5,430,799 issued to the present inventor (hereinafter termed as the “'799 Patent”).
When an MCG with very low integrated phase noise is driving the DUT, the output distortion products (CTB or CSO) on a particular channel generated by the nonlinear DUT subject to the unmodulated coherent multicarrier signal are CW signals having constant amplitudes that fall exactly on the channel frequency or exactly at offsets that are integer multiples of ±1.2625 MHz from the channel frequency. For a particular distortion product, one can picture the hundreds or thousands of distortion terms
Broadband Innovations, Inc.
Metjahic Safet
Strawbrich Robert C.
Sundaram T. R.
LandOfFree
Measuring composite distortion using a coherent multicarrier... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Measuring composite distortion using a coherent multicarrier..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Measuring composite distortion using a coherent multicarrier... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3230972