Agere Systems Inc.

Lsi Corporation

Agere Systems Guardian Corp.

Xingdong Dai

Max Jay Olsen

Geoffrey Zhang

Lane Allen Smith

Yasser Ahmed

Dwight David Daugherty

Mohammad Shafiul Mobin

Vladimir Sindalovsky

Richard Muscavage

Laurence Edward Bays

Dennis A Brooks

Christopher John Abel

Shawn Michael Logan

Weiwei Mao

Gary Schiessler

Eric Wilcox

Dennis Farley

Scott A Werner

Methods and apparatus are provided for counter-based digital frequency lock detection. A counter-based digital frequency lock detector in accordance with the present invention comprises a reference counter clocked by a reference clock and a target counter clocked by a target clock. The target counter is n bits and n is less than a number of bits of the reference counter. A frequency offset violation of the target clock is detected by comparing a value of the target counter to an n bit counter.

Method and apparatus for continuous-averaging counter-based digital frequency lock detector

In order to compensate for phase offset between different sets of circuitry having different synchronous clock domains, transmit (TX) circuitry of one domain is configured to transmit a pattern signal (e.g., a pseudo random bit sequence) to receive (RX) circuitry of the other domain. The RX circuitry cycles through a number of different phase-shifted RX clock signals to determine which selected clock signals result in valid RX pattern signals. The RX circuitry is then able to select one of the phase-shifted clock signals for use in normal processing of an RX data signal received from the TX circuitry.

Phase alignment between phase-skewed clock domains

Methods and apparatus are provided for detecting and decoding adaptive equalization training frames (having a frame marker comprised of a string of binary ones and binary zeroes). Training frames are detected by shifting the received data; inserting at least one binary value at one end of the shifted received data to generate a modified version of the received data; applying a logic function to the received data and the modified version of the received data that identifies when corresponding bit positions have different values; and detecting the frame marker when an output of the logic function has a first binary value in an approximate middle of a string of a second binary value. The training frames are decoded using a distance between the approximate center of the frame marker and a predefined binary value in an output of the logic function.

Methods and apparatus for detecting and decoding adaptive equalization training frames

Communication system optimization using a soft receiver masking technique is disclosed. For example, a method for testing a communication device comprises obtaining a software representation of a receiver portion of a given communication device. A data signal received from a transmitter through a given link channel is then processed, wherein the processing step is performed using the software representation of the receiver portion of the communication device. An output signal is caused to be generated by the software representation of the receiver portion. An eye mask test is then applied to the output signal. Based on a result of the eye mask test, one or more parameters of the transmitter may be adjusted.

System optimization using soft receiver masking technique

Techniques are disclosed for asynchronous calibration for eye diagram generation. For example, a method for calibrating a process for generating a data eye associated with a received signal comprises the following steps. Samples of the received signal are obtained for a first unit interval using a first data latch and a roaming latch. A delay offset is determined between the first data latch and the roaming latch by comparing at least one sample obtained using the first data latch and at least one sample obtained using the roaming latch, wherein the delay offset determined by the comparison is used to calibrate the process for generating the data eye associated with the received signal. A similar comparison may be done for a second data latch and used to calibrate the process. The method is able to find the accurate position of each data latch with respect to the roaming latch so as to improve the accuracy of data decoding in a digital receiver, i.e., provide receiver optimization.

Asynchronous calibration for eye diagram generation

Methods and apparatus are provided for joint adaptation of filter values in two communicating devices, such as a link partner and a link device. The disclosed joint adaptation process initially adapts the filter coefficient values in a first of the two communicating devices until a predefined stopping criteria is satisfied. Thereafter, the filter coefficient values in a second of the two communicating devices are adapted once the predefined stopping criteria for the first communicating device is satisfied. The filter coefficient values can comprise coefficient values of a multi-tap filter. The predefined stopping criteria may determine, for example, whether the first of the two communicating devices is overequalized. The filter coefficient values can be determined by including a contribution of only certain cursor tap values of the channel impulse response.

Methods and apparatus for joint adaptation of transmitter transversal filter in communication devices

Methods and apparatus are provided for detecting and decoding adaptive equalization training frames (having a frame marker comprised of a string of binary ones and binary zeroes). Training frames are detected by shifting the received data; inserting at least one binary value at one end of the shifted received data to generate a modified version of the received data; applying a logic function to the received data and the modified version of the received data that identifies when corresponding bit positions have different values; and detecting the frame marker when an output of the logic function has a first binary value in an approximate middle of a string of a second binary value. The training frames are decoded using a distance between the approximate center of the frame maker and a predefined binary value in an output of the logic function.

In serial communications, jitter is an unwanted variation of one or more signal characteristics. Two-dimensional modulation circuits and methods incorporate an amplitude pre-emphasis scheme as well as a transmit duty cycle pre-distortion (pre-DCD) technique to reduce jitter. The pre-DCD technique directly addresses transition edges of the data signal and is combined with amplitude pre-emphasis to improved data transmission.

Method and apparatus for duty cycle pre-distortion and two-dimensional modulation

A method and system for modeling and calibrating duty cycle distortion (DCD) of a Serializer and Deserializer (SerDes) device, including first generating a clock DCD signal. Once the clock DCD signal is generated, it is calibrating based upon results obtained from a filtering process of the clock DCD signal. Once the clock DCD signal is calibrated, a data DCD signal is generated and calibrated based upon results obtained from a filtering process of the data DCD signal.

Duty cycle distortion (DCD) jitter modeling, calibration and generation methods

In described embodiments, an adaptive equalizer employed by a receiver in a communication channel, such as Fibre Channel, employs pattern recognition. When a repeating pattern, such as an IDLE or ARBFF pattern, is employed by a standard to, for example, maintain a communication link, an equalizer of the receiver might adaptively set its equalizer parameters based on characteristics of the signal energy of the repeating pattern rather than adaptively set its equalizer parameters based on characteristics of the signal energy of generally random user data carried on the link. Pattern recognition by the receiver allows for maintaining adaptive equalizer parameters at settings preferred for data detection of the typical random data, improving data detection performance of the receiver when the channel transitions from a preset or synchronization repeating pattern to a user random data pattern.

Adaptive equalization employing pattern recognition

In described embodiments, an adaptive equalizer employed by a receiver in a communication channel, such as Fiber Channel, employs pattern recognition. When a repeating pattern, such as an IDLE or ARBFF pattern, is employed by a standard to, for example, maintain a communication link, an equalizer of the receiver might adaptively set its equalizer parameters based on characteristics of the signal energy of the repeating pattern rather than adaptively set its equalizer parameters based on characteristics of the signal energy of generally random user data carried on the link. Pattern recognition by the receiver allows for maintaining adaptive equalizer parameters at settings preferred for data detection of the typical random data, improving data detection performance of the receiver when the channel transitions from a preset or synchronization repeating pattern to a user random data pattern.

In an exemplary embodiment, a data alignment system comprises a First-In First-Out register (FIFO), a programmable pattern generator connected to the FIFO, and a controller connected to the programmable pattern generator and the FIFO. The FIFO is configured to provide data to or receive data from a first data lane of a serial data link having one or more lanes. Each data lane of the serial data link is configured to transmit a respective serial data stream. The programmable pattern generator is configured to generate a plurality of alignment symbols. The controller is configured to manage the alignment of the one or more data lanes of the serial data link and the insertion of a selected one of the plurality of alignment symbols into each of the serial data streams.

Data alignment method for arbitrary input with programmable content deskewing info

The present invention implements a method and apparatus for using components within a Serializer/DeSerializer (SerDes) to emulate the effects of a backplane in order to facilitate automated test equipment (ATE) testing of the SerDes. The SerDes includes a transmitter pre-emphasis circuit (TPXE) that pre-emphasizes a transmitted signal and a receiver equalization circuit (RXEQ) that equalizes a received signal. The TPXE includes coefficients that are dynamically programmable.

Backplane emulation technique for automated testing

The present invention implements an apparatus for calibrating a phase locked loop (PLL) circuit. The apparatus includes a detector for detecting frequencies of a reference signal and a controlled oscillator contained in the PLL circuit. The detector outputs the frequency difference to a control circuit. The control circuit is programmed to adjust one or more control signals to the controlled oscillator based upon the frequency difference in an orderly fashion to complete the calibration process.

Programmable linear trimming method and system for phase locked loop circuit calibration

Methods and apparatus are provided for automatic rate identification and channel synchronization in a master-slave setting for high data throughput applications. An interface is provided for use between a parallel bus and a serial bus. The interface includes a plurality of serializer/deserializer circuits that generate a clock signal, wherein one of the serializer/deserializer circuits is a master circuit generating a master clock signal and the remaining of the serializer/deserializer circuits are slave circuits generating slave clock signals. The master clock signal is substantially phase-aligned to a reference clock and is distributed to the slave circuits. The interface also includes a clock divider associated with the master circuit for selectively generating a master clock signal having one or more lower data rates than the reference clock; and a frequency detector associated with each of the slave circuits for automatically detecting a rate of the master clock signal.

Growing community of inventors

Bethlehem, PA, United States of America

Xingdong Dai