Pulse or digital communications – Equalizers – Automatic
Reexamination Certificate
1998-11-23
2002-08-20
Ghayour, Mohammad H. (Department: 2734)
Pulse or digital communications
Equalizers
Automatic
C375S229000, C375S230000
Reexamination Certificate
active
06438162
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to the field of digital transmission, and in particular to an adaptive equalizer.
BACKGROUND TO THE INVENTION
It is well known that digital pulses transmitted through telephone lines become attenuated and very distorted. Both the attenuation and the distortion are nonlinear functions of both length of the telephone line and the transmission frequency. Equalizers are used to restore the pulses to their original amplitude and shape, but introduce some jitter.
General approaches to solve the attenuation and distortion problems utilize a peak detector of one type or another to measure the amplitude of the incoming signal and compare it with some reference. Reconstitution can be effected since the amplitude and shape of the original signal at the transmitter is known. The comparing circuit generates a feedback error signal which is proportional to the loss incurred in the telephone line. The error signal is used to adjust the transfer function of the equalizer until the recovered signal generates no error signal. An edge detector (eye opening monitor) can be used to help a decision circuit create a more accurate error signal.
Known solutions are based on the above principle, but implementation methods vary. Several types of implementation methods are as follows:
(a) A simple peak detector—passive equalizing network method, which has a fixed pole and variable zero combination in the complex impedance plane. The position of the zero in the plane is varied by varying the current passing through a diode, causing its resistance to change, and thus the impedance of the passive equalizing network. The tuning range of the equalizer is determined by the values of the components.
The main advantages of this method are that it is simple to implement, and the transfer function can be easily modeled.
The main drawbacks of the method are that the single pole-zero pair causes high jitter in the output signal from the equalizer. It has a limited tuning range. The method is not suitable for monolithic integrated circuit implementation. The telephone line model is hard wired in the components.
(b) A variable pole—variable zero method, which uses a pole-zero pair in which the positions of both the pole and the zero are variable. A peak detector generates a control voltage that drives field effect transistor (FET) gates, which form nonlinear resistance elements for both the pole and the zero circuits. The model in this case is a hyperbolic function which represents a close approximation of the telephone line transfer function.
The main advantage of this method is that it is simple to implement.
The main disadvantages of this method is its limited tuning range, and that the components of the model are hard wired.
(c) A programmable automatic gain control (AGC) circuit—peak detector combination method, in which only the amplitude of the signal is recovered. The transfer function is set by selecting one of a fixed number of possible settings.
The disadvantages of this method are its low flexibility, its limited tuning range, and that the components of the model are hard wired.
(d) A switched capacitor method, which substantially eliminates the need for a passive network to create a nonlinear transfer function.
The advantage of this method is that it can be implemented in a monolithic integrated circuit.
The main disadvantages of the method are that it is inherently low speed, and that there is need for a special clock signal for sampling, which clock speed is much higher than the data rate. In addition, the components of the model are hard wired.
(e) The capacitor array method, which is based on an array of capacitors which have sizes set in a monotonically increasing order. By turning capacitors on and off, one can change the shape of the transfer function, and therefore create an adaptive equalizer.
The main advantages of this method are that it can be implemented in a monolithic integrated circuit, and that digital control feedback is possible.
The main disadvantages of this method are that it has low flexibility, and there is a need for several operational regions due to the components having limited tuning range. In addition, the components of the model are hard wired.
A description of prior art methods may be found in U.S. Pat. Nos. 3,568,100, 5,257,286, 4,606,043, 4,745,622, 4,887,278 and 5,627,885.
SUMMARY OF THE INVENTION
The present invention eliminates substantially all of the drawbacks of the aforenoted prior art, and at the same time may be fabricated using monolithic integrated circuit technology. The present invention is easy and fast to develop for a particular application, and has a flexible and programmable nonlinear transfer curve which is programmable even by the customer of the equipment in which it resides. It also has an open loop mode which can be used to debug and tune the circuit. It is suitable for fully monolithic implementation, and for high-speed applications such as T
1
and E
1
type communication systems. It also requires minimal configuration by the user, e.g. a minimal number of programmable operational regions (long and short haul).
In accordance with an embodiment of the invention, a method of equalizing a signal degraded as a result of passing through a transmission medium, comprises:
(a) storing plural equalizer transfer function control values in a memory,
(b) passing the signal through an equalizer having a controllable transfer function,
(c) comparing a characteristic of the output signal of the equalizer with a reference signal and producing a difference signal,
(d) using the difference signal to select a set of stored transfer function control values from the memory,
(e) controlling the equalizer from the selected transfer function control values so as to minimize their difference from the reference signal.
In accordance with another embodiment, an adaptive equalizer apparatus comprises:
(a) an equalizer for a signal received from a transmission medium, the equalizer having control inputs for receiving control signals for controlling the transfer function of the equalizer,
(b) a peak detector for detecting a peak amplitude of an output signal of the equalizer,
(c) a memory for storing plural values of transmission medium characteristics for plural points along the transfer function,
(d) a selection control circuit for comparing the peak with a reference and for causing selection of transmission medium characteristic values at the plural points so as to have minimum difference of the peak amplitude from the reference, and
(e) an equalizer control circuit for receiving the selected transmission medium characteristic values from the memory and for applying the control signals for controlling the transfer characteristics of the equalizer, to the control inputs of the equalizer.
REFERENCES:
patent: 4459698 (1984-07-01), Yumoto et al.
patent: 4472733 (1984-09-01), Bolger
patent: 4580275 (1986-04-01), Pirani et al.
patent: 4707840 (1987-11-01), Nakayama
Boyd Graeme B.
Culley David Heath
Sobot Robert
Baker Harold C.
Ghayour Mohammad H.
Hendry Robert G.
PMC-Sierra Ltd.
Wilkes Robert A.
LandOfFree
Implementation method for adaptive equalizer in CMOS does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Implementation method for adaptive equalizer in CMOS, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Implementation method for adaptive equalizer in CMOS will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2973351