Legerity, Inc.

Advanced Micro Devices Corporation

Muoi V Huynh

Eugen Gershon

Chien-Meen Hwang

Maged F Barsoum

Hungming Chang

Fred Berkowitz

Bin Guo

David M Barton

A network node configured for transmitting and receiving data to and from other network nodes is able to adapt the transmission rate based on the network conditions. The node initially transmits the data to a receiving node at a first rate. If the data is not received error-free, the node is able to reduce the number of data bits of the current packet that are being transmitted and to increase the amount of redundant data. The node repeats the process until error-free transmission is obtained.

Adaptive transmission system in a network

A random-access local network having multiple nodes provides data communication across residential wiring such as telephone line as a network medium, where each node accesses the network medium using discrete multi-tone (DMT) modulated symbols. The effects of amplitude and phase distortion of transmitted DMT symbols are overcome, without the necessity of complex equalizers, by differentially encoding data prior to transmission, and recovering the transmitted data by comparing phase differentials between consecutive symbol tones. Each transmitted symbol is composed of a plurality of tone signals, each tone signal modulated according to a constellation point in a complex domain. A transmitter in a first node transmits a training symbol onto a network medium, where the DMT tones of the training symbol have an equal predetermined amount of energy. A second node, having received the training symbol from the network medium, transmits the received symbol as a retransmitted training symbol back onto the network medium. The first node receives the retransmitted training symbol and checks the energy level of each DMT tone of the retransmitted training symbol. Hence, the first node can identify certain DMT tones of the retransmitted training symbol as highly attenuated if the respective energy levels fall below a prescribed threshold. The first node then selects the DMT tones for transmission by avoiding the certain DMT tones that are highly attenuated.

Methods and apparatus for channel adaptation in a DMT based system

A physical layer device (PHY) device in a home LAN employs discrete multitone technology (DMT). The DMT system enables usage of existing residential wiring, which typically is telephone system grade twisted copper pair. The PHY device comprises a spectral image selection circuit to increase frequency diversity by examining all the images captured by a filter and restoring the transmitted signal based upon an image that may not be associated with the carrier frequency but its harmonics. Alternatively, the PHY device adjusts the sampling rate of its digital to analog convertor and accordingly filters to achieve desired grouping of spectral images.

Method and apparatus for achieving frequency diversity by use of multiple images

A physical layer device (PHY) device in a home LAN employs discrete multitone technology (DMT). The DMT system enables usage of existing residential wiring, which typically is telephone system grade twisted copper pair. The PHY device comprises an analog front end (AFE) circuit that frequency shifts the spectral images by using a clock signal. The multiplication of the clock signal is accomplished using a digital mixer or, in the alternative, an analog switch.

Method and apparatus for frequency shifting with a clock signal

A random access data network in accord with the invention will utilize discrete multi-tone (DMT) transmission over a link within a home or other premises, such as over existing in-house telephone wiring. Discrete multi-tone communication in a random access implementation raises a number of timing issues. To address these issues, all packet transmissions through the network will begin with one or more cycles of a defined periodic signal, typically a sinusoid at a specified frequency. All timing of decoding of subsequent signals, such as ID signals and DMT symbols, is based on counting signal sampling intervals following the detected timing mark. The preferred embodiment utilizes differential modulation of the tones within the DMT symbol. Each packet may include a reference symbol to enable demodulation of the first DMT data symbol. Also, the preferred embodiment of the DMT symbol includes a prefix code followed by the actual DMT waveform carrying the modulated information. The prefix code corresponds to a fixed number of samples of time-domain data, such as a copy of the last thirty-two samples used to generate the actual DMT waveform.

Timing recovery scheme for a discrete multitone transmission system

A random-access local network having multiple nodes provides data communication across residential wiring such as telephone line as a network medium, where each node accesses the network medium using discrete multi-tone (DMT) modulated symbols. The effects of amplitude and phase distortion of transmitted DMT symbols are overcome, without the necessity of complex equalizers, by differentially encoding data prior to transmission, and recovering the transmitted data by comparing phase differentials between consecutive symbol tones. Each transmitted symbol is composed of a plurality of tone signals, each tone signal modulated according to a constellation point in a complex domain. A transmitter modulates data, such as a bit-pair, to a new constellation point based on the value of the bit-pair and a prior position of a consecutively-preceding constellation point, such that the data is represented by the difference in positions between the consecutively-preceding constellation point and the new constellation point. The constellation points are then used to modulate a selected one of the tone signals. Although a receiver detecting the modulated tone signal may recover the transmitted constellation points at different absolute positions due to distortion on he network medium, the relative difference between the consecutively-preceding constellation point and the new constellation point is preserved, enabling the receiver to recover the transmitted data based on the relative difference in positions of the consecutive constellation points. Hence, entire bit stream can reliably transmitted using position changes between consecutive constellation points.

Differential encoding arrangement for a discrete multi-tone transmission system

A novel collision detecting method at the physical layer is provided for use in a multi-point DMT communication system. A time mark transmitted at the beginning of a portion of a random access data communication is followed by an ID signal for identifying the communication station. The ID signal produced based on the time mark may be transmitted between the time mark and a data portion of the random access data communication. The time mark distorted by a communication channel is received by a receiver of the station that transmitted it. An additional ID signal representing the same ID data as the ID signal is produced based on the received time mark. The additional ID signal is compared with the ID signal received from the communication channel. A collision event is detected if the incoming ID signal does not match the additional ID signal. In response to the detection of the collision event, the transmission of the data portion is terminated.

Collision handling scheme for discrete multi-tone data communications network

A method and apparatus for detecting a sinusoidal signal samples a received signal. An error signal generator receives at its inputs two previous samples of the signal and a current sample of the signal and generates an error signal based on these previous and current samples. A comparison circuit compares the generated error signal for the current sample to an error threshold value and generates a threshold comparison signal that indicates whether the generated error signal is below the error threshold value. A determination circuit then determines whether the received signal is a sinusoidal signal based on the threshold comparison signals that are generated for a plurality of samples.

Method and apparatus for detecting a sinusoidal signal

A method and apparatus for detecting a sinusoidal signal samples a received signal. An error signal generator receives at its inputs a previous sample of the received signal and a current sample of the received signal and generates an error signal based on these previous and current samples. A comparison circuit compares the generated error signal for the current sample to an error threshold value and generates a threshold comparison signal with a first value that indicates the generated error signal is below the error threshold value for a second value that indicates a generated error signal is above the error threshold value. A determination circuit then determines whether the received signal is a sinusoidal signal based on a threshold comparison signal generated for a plurality of samples. The determination circuit includes a counter that maintains a count of the number of threshold comparison signals having the first value within a sampling period.

Method and apparatus for detecting a sinusoidal signal using a counter and queue

A discrete multi-tone communication system employing bin-to-bin differential encoding of data frames to be transmitted within the communication system. The bin-to-bin differential encoding utilizes the phase angle of previously encoded data to encode the current data relative to that phase angle. Hence, only a single reference tone is required to encode a first portion of the data frame into a discrete tone, and then the remaining data portions of the data frame are each subsequently encoded into discrete tones with reference to the phase angle of data already encoded into a discrete tone. Thus, to achieve decoding, a decoding device only requires the reference tone to begin decoding the discrete tones to the correct phase angle for an accurate reproduction of the original data frame.

Bin-to-bin differential encoding apparatus and method for a discrete multi-tone transmission system

A novel collision detecting method is provided at the physical layer for use in a multi-point DMT communication network having stations transmitting over common media. The method includes generating a collision code corresponding to ID information of a station. A collision symbol representing the collision code is transmitted at the beginning of a portion of a random access data communication. A station in the network receives the collision symbol and decodes a collision signal carried by the collision symbol. A collision event is detected if the collision signal does not satisfy pre-set requirements. For example, energy of each component of the collision signal may be compared with a predetermined minimum energy level to generate a code element for each component having energy exceeding the predetermined minimum energy level. Validity of the code having the generated code elements is determined based on a predetermined validity criterion. Also, the distance in frequency domain between components of the collision signal corresponding to the generated code elements is compared with a predetermined minimum distance in frequency domain, to confirm validity of the detected code.

Collision detection system for multiple stations in discrete multi-tone

A method and arrangement for transferring data between nodes over a telephone cable that couples the nodes provides a dual configuration node transmitter, and a dual configuration node receiver. The dual configuration transmitter has a first transmitter circuit that transmits data over a first configuration of the telephone cable and a second transmitter that simultaneously transmits other data over a second configuration of the telephone cable. The dual configuration receiver is coupled to the telephone cable and has a first receiver circuit that selects the data transmitted by the first transmitter circuit and rejects the data transmitted by the second transmitter circuit, and a second receiver circuit that selects the data transmitted by the second transmitter circuit and rejects the data transmitted by the first transmitter circuit. The transmitting of data over two configurations of the same cable allows data to be transmitted twice as fast as over the same cable set up for transmission in a single configuration. Alternatively, in an Ethernet application, one configuration may carry the data while the other configuration simultaneously carries activity information.

Growing community of inventors

San Jose, CA, United States of America

Muoi V Huynh