Amplifiers – With control of power supply or bias voltage – With control of input electrode or gain control electrode bias
Reexamination Certificate
2000-12-22
2003-09-16
Nguyen, Patricia (Department: 2817)
Amplifiers
With control of power supply or bias voltage
With control of input electrode or gain control electrode bias
C327S348000, C327S309000
Reexamination Certificate
active
06621338
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to systems used for computing the correlation between two signals and, more particularly, to methods of setting the gain of amplifiers used in such systems.
BACKGROUND OF THE INVENTION
Digital transmission systems for communication usually include links in which many slower data streams are multiplexed together onto a single higher speed link. This higher speed link may be implemented by an optical transmission system. Such systems can be end-to-end links, or networks of optical links, with switching elements. Testing and verifying the operation of switching elements of such an optical network can pose problems, particularly because the data rates are often very high.
It is known that connectivity through a network can be guaranteed by network layer protocols such as TCP/IP, which can request retransmission if some data does not arrive. At the physical data transfer level, path trace bytes can be inserted and checked to ensure that a frame of data is routed correctly. However, such methods require decoding the data stream, which usually only takes place at a terminal at the end of the path through the optical network.
With a view to performing connection verification in a non-intrusive manner at an intermediate switching element, developments have been made in the field of testing using pattern matching. For example, as disclosed in U.S. Pat. No. 6,005,695 issued to Roberts, assigned to the assignee of the present invention and incorporated by reference herein, pattern matching can be carried out by a correlation process involving the low frequency components of an input signal and a switched signal. Since the connection map being applied by the switching element is known, it is a simple matter to determine the expected result (i.e., low or high) of the correlation process. By performing a comparison of the measured and expected correlation values, one can assess whether the switching element is operating properly or is malfunctioning.
It should therefore be apparent that obtaining a reliable correlation value from the correlation process is essential to making a correct decision as to whether or not the switching element is operating correctly. Since the input signal and the switched signal often have different signal levels, for example due to losses suffered within the switching element, it is important to provide appropriate amplification or attenuation (i.e., “level setting”) prior to computing the correlation value. One way of providing level setting relies on a pair of programmable-gain amplifiers, one on the input signal side and another one on the switched signal side. A significant challenge lies in appropriately setting the gains of the two amplifiers such that the ensuing correlation results are reliable.
SUMMARY OF THE INVENTION
It has been found that when clipping occurs at least part of the time, the correlation result obtained for a pair of truly correlated signals remains virtually unaffected, while there is actually a beneficial effect on the correlation result obtained for a pair of truly uncorrelated signals. Accordingly, the present invention provides for each signal to be amplified to such an extent that the amplified version of the signal will be clipped by a subsequent analog-to-digital converter for a desired percentage of the time, on average. A correlation value is then computed in the digital domain based on the clipped signals. To determine the amplifier gain needed to attain the desired amount of clipping, the present invention capitalizes on the availability of power measurements taken from the signal before it was amplified. In this way, reliance on feedback from the amplifier output is not required, thus simplifying the gain computation process.
Accordingly, the invention may be summarized broadly as a method of setting the level of an analog signal supplied to an electronic device having a maximum input level, including computing a gain which, when applied to the analog signal, causes the resulting amplified signal to exceed the maximum input level for a percentage of time which is greater than zero, and applying the gain to the analog signal.
The invention may also be broadly summarized as a combination which includes an electronic device having an input terminal and characterized by a maximum input level, and an amplifier connected to the input terminal of the electronic device, for applying a gain to an analog input signal, the gain being selected such that the percentage of time during which the signal supplied to the electronic device exceeds the maximum input level is greater than zero.
The invention may also be broadly summarized as a method of jointly processing a first signal and a second signal, including applying a first gain to the first signal and applying a second gain to the second signal, thereby to produce first and second amplified signals, respectively; limiting the amplitude of the first amplified signal to within a first fixed range and limiting the amplitude of the second amplified signal to within a second fixed range; digitally correlating the amplitude-limited signals; and controlling the first gain so as to cause the first amplified signal to fall outside the first fixed range during a first percentage of time that is greater than zero and controlling the second gain so as to cause the second amplified signal to fall outside the second fixed range during a second percentage of time that is greater than zero.
The invention may further be broadly summarized as a system for processing a pair of analog signals, including a first unit for applying a gain to each signal; a second unit for limiting the amplitude of each amplified signal to within a maximum range; a third unit for digitally correlating the amplitude-limited signals; and a fourth unit for controlling the gain applied to each signal so as to cause the corresponding amplified signal to fall outside the maximum range during a percentage of the time that is greater than zero.
In addition, the invention may be broadly summarized as a correlation engine, including a first analog-to-digital converter (ADC) having a fixed input range, a first programmable-gain amplifier connected to an input of the first ADC, a second ADC having a fixed input range, a second programmable-gain amplifier connected to an input of the second ADC, a correlator connected to the first and second ADCs, and a control element connected to the first and second amplifiers, for setting the gain of the first and second amplifiers to such a value that the average percentage of time during which the signal amplified by the each amplifier falls outside the input range of the respective ADC is greater than zero.
The invention may also be broadly summarized as computer-readable media tangibly embodying a program of instructions executable by a computer to perform a method of computing a gain of an amplifier used for amplifying an analog signal to be fed to an electronic device having a maximum input level V
MAX
, wherein the analog signal is the result of low-pass filtering an F
C
-bps bit stream to F
B
hertz, the method including continuously reading a mean level V
AVG
of the analog signal and computing the gain as substantially equal to the quotient of (V
MAX
×V
AVG
×(2 F
LP
)
½
) and (&bgr;×(R
CLK
)
½
), where &bgr; is chosen such that the value of the complementary error function (erfc) at &bgr; corresponds to a desired probability of clipping.
REFERENCES:
patent: 5027117 (1991-06-01), Yoshida et al.
patent: 5046107 (1991-09-01), Iwamatsu
patent: 5239693 (1993-08-01), Gailus et al.
patent: 5459426 (1995-10-01), Hori
patent: 5659313 (1997-08-01), Dischert et al.
patent: 5752171 (1998-05-01), Akiya
patent: 5910751 (1999-06-01), Winn et al.
patent: 6047167 (2000-04-01), Yamashita
patent: 6195029 (2001-02-01), Urry et al.
patent: 6236864 (2001-05-01), McGowan et al.
Nguyen Patricia
Nortel Networks Limited
LandOfFree
Gain determination for correlation processes does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Gain determination for correlation processes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Gain determination for correlation processes will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3084655