Pulse or digital communications – Synchronizers – Synchronizing the sampling time of digital data
Reexamination Certificate
2000-07-06
2001-07-31
Pham, Chi (Department: 2631)
Pulse or digital communications
Synchronizers
Synchronizing the sampling time of digital data
C375S371000, C370S517000, C359S199200
Reexamination Certificate
active
06269137
ABSTRACT:
BACKGROUND OF THE INVENTION
Many communications networks today provide high bit-rate transport over a shared medium, such as passive optical networks (PON), cable television (CATV) coaxial or hybrid fiber/coax networks, and wireless. These shared medium networks typically use time, frequency or code division multiplexing to transport data signals from a central terminal to several remote customer terminals and time division multiple access (TDMA) to transport data signals from the customer terminals to the central terminal. TDMA is characterized by non-continuous, or burst mode, data transmission.
Traditional clock and data recovery methods are generally optimized for communications systems which receive continuous data streams that have enriched spectra at the sampling frequency. In the case of non-continuous (burst mode) data transmission, the use of a very long preamble or an embedded clock is typically required to allow for data recovery when using traditional clock/data recovery methods. However, both approaches are wasteful of the link bandwidth and are not suited to making very fast (e.g., <20 ns) data recovery decisions.
Other recovery methods sample a known preamble using multiple clock signals having different phases.
SUMMARY OF THE INVENTION
It is desirable to be able to quickly (e.g., <20 ns) recover data from burst mode transmissions without wasting link bandwidth.
Rather than adjusting the local clock to coincide with the data, the approach of the present method and apparatus selects from N incrementally delayed copies of an incoming data stream the data stream that is most aligned with a local clock signal. That is, the local clock is kept fixed and the data is phase and byte aligned to the local clock. The present approach optimizes alignment of the rising edge of the local clock with the center of the bit-time of the data stream.
By splitting the incoming data stream into N identical streams and then delaying each of the streams by multiples of a bit-time/N, several sampling choices are presented simultaneously. These delayed streams are fed into high speed sampling circuits each clocked with the same local oscillator. A logical decode of the now parallel sampled streams is made to determine from at least two bit transitions which of the data streams is the best choice in relation to the local oscillator for recovering the data. The selected data stream is thereby retimed to the local oscillator for synchronous processing of the payload.
With the present approach link bandwidth is maximized by avoiding the need for a long preamble or an embedded clock, data is recovered rapidly (within 20 ns), and the local system clock is held constant in the presence of multiple asynchronous data streams for improved robustness in overall system performance.
Accordingly, a method of data recovery includes receiving a data stream of data bits and splitting the data stream to N identical input data streams where N is an integer greater than 1. Each of the N input data streams is delayed with respect to the preceding one by a bit time divided by N. Each of the N delayed input data streams is sampled using a local clock to provide N samples which form an N-bit sample code per clock period. At least two successive sample codes are decoded to select one of the N delayed input data streams most aligned with the local clock.
According to an aspect of the method, the received data stream comprises a burst transmission which includes at least two bit transitions which are sampled.
According to another aspect of the method, the decoding includes declaring a sample code valid if the sample code includes at least X consecutive zeroes or ones and if the sample code or adjacent sample codes include at most Y consecutive zeroes and ones, otherwise a sample code is declared invalid. At least two successive valid sample codes are decoded to select the data stream most aligned with the local clock.
According to another aspect, data recovery in a system in which a central terminal and plural remote terminals communicate over a shared medium network includes receiving at the central terminal a data stream of data bits transmitted from one of the plural remote terminals over the shared median network. The data stream is split to plural identical input data streams each of which is delayed with respect to the preceding one by a time interval less than a bit time. Each of the delayed input data streams is sampled using a local clock to provide samples that form a sample code per clock period. At least two subsequent sample codes are decoded to select one of the plural delayed input data streams most aligned with the local clock.
In accordance with the invention, a data recovery circuit for recovering data bits from a received data stream includes a splitter for splitting the data stream to N identical input data streams and N delay lines each having an input coupled to a corresponding one of the N input data streams. Each delay line delays the corresponding input data stream with respect to the preceding one by a bit time divided by N. The data recovery circuit further includes N samplers each having an input coupled to an output of a corresponding one of the N delay lines. Each sampler samples the corresponding delayed input data stream using a local clock to provide N samples forming an N-bit sample code per clock period. A decoder/selector decodes at least two successive sample codes and selects one of the N delayed input data streams most aligned with the local clock.
The principles of the present invention can be applied to telecommunications, data communications, cable television transmission systems or any other network which uses time division multiple access to transmit narrowband, wideband and broadband services such as data, voice, video, and image.
REFERENCES:
patent: 5022056 (1991-06-01), Henderson et al.
patent: 5173899 (1992-12-01), Ballance
patent: 5327277 (1994-07-01), Van Der Plas et al.
patent: 5398129 (1995-03-01), Reimann
patent: 5479451 (1995-12-01), Eldering et al.
patent: 5488639 (1996-01-01), MacWilliams et al.
patent: 5572349 (1996-11-01), Hale et al.
patent: 5680546 (1997-10-01), Chiaretti et al.
patent: 5754535 (1998-05-01), Vandenabeele et al.
patent: 5774244 (1998-06-01), Tandon et al.
patent: 5818890 (1998-10-01), Ford et al.
patent: 5822386 (1998-10-01), Pawelski
patent: 5848070 (1998-12-01), Durvaux et al.
patent: 5864413 (1999-01-01), Feldman et al.
patent: 5864414 (1999-01-01), Barnsley et al.
patent: 5872645 (1999-02-01), Proctor
patent: 5896213 (1999-04-01), Nagahori et al.
patent: 5896474 (1999-04-01), Van Deventer et al.
patent: 5907587 (1999-05-01), Sokoler
patent: 5926478 (1999-07-01), Ghaibeh et al.
patent: 5930018 (1999-07-01), Effenberger
patent: 5930262 (1999-07-01), Sierens et al.
patent: 6122335 (2000-09-01), Colella et al.
patent: 0 889 611 A1 (1999-01-01), None
Denis J.G. Mestdagh; Fundamentals of Multiaccess Optical Fiber Networks; ISBN 0-89006; pp. 332-337; 1995.
ITU-T Recommendation G.983.1 “Broadbend Optical Access System Based on Passive Optical Networks (PON)”, Oct. 1998.
Box Brian
Colella Barry D.
Farrar Lewis W.
Masucci Jeffrey A.
Hamilton Brook Smith & Reynolds P.C.
Pham Chi
Phu Phuong
Quantum Bridge Communications Inc.
LandOfFree
Method and apparatus for fast burst mode data recovery does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for fast burst mode data recovery, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for fast burst mode data recovery will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2466231