WO1997034390A1

WO1997034390A1 - High speed data communication over unshielded twisted pair

Info

Publication number: WO1997034390A1
Application number: PCT/US1997/003906
Authority: WO
Inventors: Jacques G. Ruch; David B. Ribner
Original assignee: Analog Devices, Inc.
Priority date: 1996-03-13
Filing date: 1997-03-13
Publication date: 1997-09-18

Abstract

A duplex modulation method including the steps of using a first modulation technique to send downstream information over a cable and within a first region of the signal frequency spectrum; and using a multicarrier modulation technique to send upstream information over the cable and within a second region of the frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, wherein the first modulation technique is different from the multicarrier modulation technique.

Description

High speed data communication over unshielded twisted pair

Background of the Invention

The invention relates generally to asymmetric digital subscriber lines (ADSL) and switched digital video (SDV) networks.

Fig. 1 shows a typical fiber-to-the-curb (FTTC) architecture such as might be used to deliver switched digital video signals (also referred to as downstream signals) to customers' homes. This is an architecture that telephone companies use to offer broadband services to residential customers. The goal is to offer both TV services (e.g. switched digital video services) , data services (e.g. Internet access) , and plain old telephone service (POTS) to the residential customers. In the FTTC system, customers are provided with equipment which enables them to receive and decode signals from the telephone company and to send signals back to the telephone company, e.g. signals which are used to electronically select the particular signals that they wish to have delivered to them. In general, a telephone central office (CO) 10 delivers both telephone signals and the switched digital video signals to customers and a CATV source (e.g. a cable head end) 12 delivers cable TV signals to the customers over the same lines within the customer' s residence. These two sources might be collocated or they might be located remotely from each other. On ordinary cable TV, one already has all of the TV channels above 50 MHz, each located in its own 6 MHz slot. The switched video comes from the telephone part of the connection and is located below 50 MHz. The video signals at CO 10 are converted to digital, compressed using a standard algorithm (e.g. MPEG) , and then sent to an optical network unit (ONU) 16, which is typically mounted on a telephone pole located near the residences which are being served (e.g. 1500 feet) . Each ONU 16 services at most 16 residences. ONU 16 remodulates the signal from CO 10 and places it in a band from 0-50 MHz, i.e., below the conventional cable channels. Then, ONU 16 transmits the remodulated signals over an unshielded twisted pair (UTP) cable 18 to the residence. One advantage of using ONU 16 is that it permits the companies to utilize preexisting UTP connections to customers' residences and it avoids having to bring optical fibers directly into all homes which would be a very time consuming and expensive job. There might be about 3000 feet of UTP 18 between ONU 16 and the residence to which the signals are being delivered.

At the residence, a passive combiner 20 combines the signals coming over UTP 18 (i.e., frequencies below 30 MHz) with the signals from the CATV source 12 (i.e., frequencies above 50 MHz) and puts them onto a coax 22 which goes into the house. Passive combiner 20, which is a conventional well known device, in addition to combining the signals also filters them and prevents them from wrapping around onto the other input line.

Within the house, a POTS (plain ordinary telephone signal) splitter 24 separates out the telephone signals, which are those signals below 4 kHz, and puts them onto the phone lines (e.g. UTP) within the house to which the telephones 28 are connected. Everything above 4 kHz remains on the coax and is sent to another splitter 30 for distribution to one or more set top boxes 32 within the residence. Set top boxes 32 are used to tune in the desired digital signal and convert it for display on a TV, video screen, or other display device. In general, a set top box includes a receiver that decodes the waveform and converts it back to a digital signal. Then, there are other modules which process the decoded signal appropriately. For example, there is a framing module that groups the bits according to a specific protocol and there is an MPEG decoder that converts the bits to image signals.

Set top boxes 32 also include circuitry which enable the user to send signals (referred to as upstream signals) back to central office 10.

To support applications such as video on demand, Internet access, and telecomputing, it is desirable to support downstream data rates to customers premises of approximately 50 Mbps while also providing upstream data rates of at least 2 Mbps. In an FTTC architecture this is made difficult by the final connection from the ONU to the customer¹ s premises, which is typically via a 600 feet or longer UTP cable. The UTP cable produces high loop attenuation above 25 MHz and it is vulnerable to a high levels of ingress from electromagnetic (EM) pickup at frequencies below 6 MHz from AM radio, home appliances, and lamp dimmers, just to mention a few sources.

In some implementations, downstream transmission is accomplished using 16-Quadrature Amplitude Modulation (QAM) or 16-Carrierless Amplitude Phase Modulation (CAP) . In general, CAP is a form of QAM. The primary difference is that CAP, unlike QAM, does not require an upconverter or a down converter. The signal is automatically put on a low frequency carrier, e.g. about 16 MHz. In contrast, conventional QAM uses a mixer in which a 16 MHz carrier is mixed with a based band signal. Note that both QAM are modulation techniques that are well known to persons skilled in the art and are both fully described in the literature. For example, with regard to CAP, see G.-H. Im, et al., "51.84 Mb/s 16-CAP ATM LAN Standard," IEEE Journal on Selected Areas in Communications, Vol. 13, No. 4, May, 1995, pp. 620-632.

Both of these modulation techniques have a spectral efficiency that is theoretically limited to 4 bits/sec/Hz. To support a payload of 50 Mbps, an overhead of roughly 10% and an excess bandwidth of 20% are necessary. Accordingly, a total bandwidth approaching 20 MHz is consumed in the downstream channel. Thus, the available bandwidth between 6 and 20 MHz is only sufficient for the downstream channel in a 50 Mbps VDSL system.

An existing approach to sending the upstream information uses quadrature phase-shift keying (QPSK) signaling and it sends the upstream information in a frequency band that is above the downstream signals, e.g. between 26-30 MHz.

One problem is that since UTP was originally designed for handling only voice signals, it seriously attenuates signals above 4 KHz. Coax, in comparison, has a bandwidth of about 750 MHz. In UTP the signal rolls off quickly above normal voice frequencies so that at 30 MHz the signal is attenuated significantly. Thus, any signal that is put above that frequency will have a low SNR. One consequence is that there is a serious limit to the length of UTP that can be used. UTP really works only up to distances of about 600 feet, sometimes not even that far. Yet, many customers need longer distances e.g. up to 3000 feet. Moreover, if impairments exist on the line, such as bridge taps or wire gauge changes, then it does not work at all.

Summary of the Invention In general, in one aspect, the invention is a duplex modulation method including the steps of: using a first modulation technique to send downstream information over a cable and within a first region of the signal frequency spectrum; and using a multicarrier modulation technique to send upstream information over said cable and within a second region of the frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from the multicarrier modulation technique. Preferred embodiments include the following features. The first modulation is a form of QAM or CAP, e.g. 2ⁿ-CAP, where n=4. The multicarrier modulation technique is a form of orthogonal frequency division multiplexing, e.g. DWMT. The first region of the signal frequency spectrum lies below about 6 MHz and the second region of the signal frequency spectrum lies between about 6 MHz and 30 MHz.

In general, in another aspect, the invention is a duplex modulation method which is employed over unshielded twisted pair cable (UTP) in a FTTC system, wherein downstream information is sent over the UTP in one direction to a customer device and upstream information is sent over the UTP in an opposite direction from the customer device. The method includes the steps of using a first modulation technique to send the downstream information over the UTP and within a first region of the signal frequency spectrum; and using a multicarrier modulation technique to send the upstream information over the UTP and within a second region of the frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from the multicarrier modulation technique. In general, in yet another aspect, the invention is an apparatus for transmitting downstream signals over a cable to a remote unit and for receiving upstream signals sent over the cable from the remote unit. The apparatus includes a transmitter which employs a first modulation technique to generate and send the downstream signals over the cable, wherein the downstream signals are within a first region of the signal frequency spectrum; and it also includes a receiver which employs a multicarrier demodulation technique to receive and demodulate the upstream signals, wherein the upstream signals are within a second region of the signal frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from the multicarrier modulation technique.

In general, in still another aspect, the invention is an apparatus for receiving downstream signals sent over the cable from a remote unit and for transmitting upstream signals over a cable to the remote unit. The apparatus includes a receiver which employs a first modulation technique to receive and demodulate the downstream signals, wherein the downstream signals are within a first region of the signal frequency spectrum; and it also includes a transmitter which employs a multicarrier modulation technique to generate and send the upstream signals over the cable, wherein the upstream signals are within a second region of the signal frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from the multicarrier modulation technique. With DWMT being used in the lower frequency range (i.e., below the downstream signals), there is far less attenuation and thus much better SNR. Thus, higher dimensional signaling constellations can be used to thereby get more bits/Hz (i.e., a higher data rates) as compared to prior art systems.

The modulation configuration of the invention supports duplex operation over lengths of 500-750 feet of category 3 UTP. It also supports downstream data rates in excess of 50 Mbps and upstream data rates in excess of 2 Mbps.

Other advantages and features will become apparent from the following description of the preferred embodiment and from the claims.

Brief Description of the Drawings

Fig. 1 is a block diagram of a representative FTTC (fiber-to-the-curb) network;

Fig. 2 shows a spectrum of signal allocation in accordance with the invention; Fig. 3 is a block diagram of an ONU (optical network unit) ; and

Fig. 4 is a block diagram of a set top unit of the type that might be located at a residence.

Description of the Preferred Embodiments Frecruency Spectrum Allocation:

A system which implements the invention allocates the downstream and upstream signals in the different portions of the frequency spectrum shown in Fig. 2. In a band above 6 MHz (see band 10) , one modulation technique (e.g. either QAM or CAP) is employed to send downstream signals. Below 6 MHz, a multicarrier modulation technique (e.g. Discrete Wavelet Modulation (DWMT)) is employed to send upstream signals. DWMT is a form of multicarrier modulation which is particularly efficient and can have very low sidelobe levels. In general, multicarrier modulation partitions the channel into many subchannels, each with its own associated subcarrier. One well known multicarrier modulation technique is referred to as discrete multitone transmission (i.e., DMT) or orthogonal frequency division multiplexing (OFDM) . In that case, the signal that is sent over each subchannel is generated by applying an orthogonal transformation to the data to be sent over the channel. At the receiver, the inverses of the orthogonal transformations are applied to the received signal to generate and separate out the signals that were encoded into the array of subchannels. DWMT is done in a similar manner except that a special pulse waveform is employed to perform the transformation and thereby achieve a higher level of subchannel isolation without sacrificing channel efficiency.

DWMT is also more commonly referred to as cosine modulated filter bank. For further information about

DWMT and how to implement it refer to R. David Koilpillai et al. "Cosine-Modulated FIR Filter Banks Satisfying Perfect Reconstruction" , IEEE Transactions on Signal Processing, Vol. 40, No. 4, April, 1992, pp. 770-783; and M.A. Tzannes et al., "DMT Systems, DWMT Systems and Digital Filter Banks," Proc. IEEE International Conference on Communications, May, 1994, pp. 311-314; and S.D. Sandberg et al., "Overlapped Discrete Multitone Modulation for High Speed Copper Wire Communications," IEEE Journal on Selected Areas in Communications, Vol. 13, No. 9, December, 1995, pp. 1571-1585.

In the described embodiment, a transmultiplexer in the set top boxes takes time-division multiplexed data and transforms it to frequency division multiplexed data (FDM) using 384 closely-spaced subcarriers, i.e., 8 KHz bandwidth with a 4 KHz spacing. In other words, the channel signals are overlapping. As indicated above, this can be done because DWMT uses functions to modulate the signals on the different subcarriers that are selected from an orthogonal basis set. Thus, the overlapping signal are orthogonal to each other. Since there will typically be more capacity than is usually needed, we can disable the subcarriers for which the SNR is unacceptably low, i.e., subcarrier frequencies that are subject to excessive ingress or narrow band interference. Since DWMT is very selective, in practice only a small number of carriers will have too much noise and must be eliminated. Thus, this approach produces a very robust solution to the problem of sending upstream data.

In a high, narrow-band ingress environment, the channel selection is actually performed dynamically. That is, all of the subcarriers of the DWMT signal are continuously monitored at the receiver end (i.e., at the ONU) and the subcarriers with the lowest signal-to-noise ratios due to ingress are avoided. Due to the narrow¬ band characteristics of individual subcarriers in DWMT, only a small number of subcarriers are affected by a single frequency interferer so that the loss in capacity can be minor. Also, it should be noted that subcarriers with the higher signal-to-noise ratios can support higher data rates. Therefore, with DWMT modulation, different levels of PAM (pulse amplitude modulation) modulation (e.g. 4-level, 6-level, etc.) can be employed on individual subcarriers depending on the SNR levels of the subcarriers. Using this approach, one can optimize the utilization of bandwidth by adapting the coding level to the characteristics of each of the subcarriers across the upstream band. An FTTC Implementation:

Referring to Fig. 3, an optical network unit 50 constructed in accordance with the invention includes an interface chip 52 which connects to fiber optic links 54a and 54b to and from a central office. Among other things, interface chip 52 reformats the data and directs it to the appropriate channel. The interface chip might implement, for example, the SONET (Synchronous Optical Network) protocol, which is a time division multiplex technique that is an international standard adopted by the CCITT (International Telegraph and Telephone Consultative Committee) in 1989.

The downstream signals received over the fiber optic link 54a for transmission to residences are first processed by a digital CAP transmit chip 56 and then a CAP transmit analog front end (AFE) chip 58. The functions that are performed by these chips are well known and thus will not be described in detail here. Suffice it to say that in general, transmit chip 56 modulates the signals using n-CAP modulation and AFE chip 58 converts the digital signals to analog signals and performs the appropriate signal conditioning so that the signals can be sent out over the UTP to the residences. Typically, n=2^m where m is an integer. In the described embodiment, 16-CAP is used (i.e., n=4) as it is particularly applicable to the cable environment.

The signals from the central office also include ordinary telephone signals. In ONU 50 there is a POTS subscriber line interface circuit 60 that processes the telephone signals and passes the processed signals to a three-way diplex filter 62 which combines them with the CAP signals for transmission as downstream signals over the UTP to the subscriber' s residence.

Referring to Fig. 4, near the subscriber' s residence, a passive combiner 70 combines the downstream signals with cable signals from a CATV source and delivers the combined signal to the subscriber' s residence. At the residence a POTS splitter 72 separates out the telephone signals (i.e., the signals below 4 KHz) and puts them onto telephone lines within the residence. The rest of the downstream signals go to another splitter 74 which distributes the downstream signals to multiple set top boxes 76 at the subscriber' s residence. In Fig. 4, only one set top box 76 is shown. Within each set top box, a diplex filter 78 directs the downstream signals to a CAP AFE receiver chip 80 and then to a CAP digital receiver chip 82. CAP AFE receiver chip 80 performs signal conditioning and analog to digital conversion. CAP receiver chip 82 demodulates the received signal and passes the demodulated signal to other circuitry which constructs a displayable image. The other circuitry might include, for example, a fra er 84 and an MPEG decoder 86, which performs the appropriate decompression on the signal. To send upstream signals back to the central office, set top box 76 includes a DWMT digital transmitter chip 88 followed by a DWMT transmit AFE chip 90. DWMT transmitter chip 88 modulates the upstream information signals using DWMT modulation and DWMT transmit AFE chip 90 performs D/A conversion, signal conditioning, and buffering of the signal. The output of DWMT AFE chip 90 goes to diplex filter 78 which places that signal onto the coax line that connects to splitter 74. The upstream signal from splitter 74 passes to POTS splitter 72 where it is combined with the telephone signals and passed on to passive combiner 70 which, in turn, places the upstream signal onto the UTP for transmission back to ONU 50.

Referring again to Fig. 3, at ONU 50 diplex filter 62 separates out and directs the upstream signal from the subscriber' s set top box to a DWMT receive AFE chip 64 and from there to a DWMT digital receiver chip 66.

Within ONU 50, a monitoring circuit 68 monitors the output of DWMT receiver chip 66 and measures the SNR of each of the subchannels. This can be done either sequentially for each subchannel or simultaneously for all of the subchannels. The SNR is measured by periodically by sending a reference signal from the set top box to the ONU. The receiver in the ONU, of course, knows what the reference signal is, thus it can compute the noise on each of the channels and from that determine the SNR for each of the channels. Using these SNR measurements, SNR monitor 68 then performs dynamic channel selection and reassignment and sends its selection and reassignment decisions to the set top boxes 76 (see Fig. 4) . An operations channel is provided within the bandwidth of the downstream signals to communicate messages to the set top boxes informing them of what subcarriers have been enabled/disabled. A controller 100 within the set top box switches the DWMT transmitter to the appropriate channels.

Protocols for selecting subcarriers, for communicating those selection decisions, and for switching to the new subcarriers are known in the art and thus are not described here. Indeed, one such SNR measurement technique and switching protocol is specified by the industry adopted standards which apply to ADSL. For a further details about a technique which can be employed to perform dynamic subchannel, the reader is referred to U.S. 5,400,322 to Hunt et al. which describes bit swapping. One possible implementation of subchannel selection can be viewed as simply a subset of the bit swapping algorithm described therein. In the case of subchannel selection, however, the decision is either to use the subcarrier or not use it; whereas in the case of bit swapping the decision is made as to how many bits to use send with each subcarrier. Of course, a more sophisticated approach to subchannel selection could also be implemented in which the bits/Hz is varied for each subcarrier depending on the measured SNR of the subchannel (i.e., the number of bits assigned to the subcarrier can be a variable parameter chosen on the basis of the SNR of the subcarrier channel) .

Operation and implementation details of the above- mentioned modules and/or components in ONU 50 and set top box 76, as well as, of the splitters are well know to persons skilled in the art. Thus, further details about those components are not presented here. Rather the reader is referred to the relevant publicly available literature some of which has been identified above. Also, it should be understood that the modules and/or components can be implemented in actual hardwired, dedicated chips as described herein or by using discrete component circuitry or some combination of discrete component circuitry and dedicated chips, or through software running on a digital signal processor or a computer. The manner in which the functionality is implemented is not of central importance.

Other embodiments are within the following claims.

What is claimed is:

Claims

Claims :

1. A duplex modulation method comprising: using a first modulation technique to send downstream information over a cable and within a first region of the signal frequency spectrum; and using a multicarrier modulation technique to send upstream information over said cable and within a second region of the frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from said multicarrier modulation technique.

2. The method of claim 1 wherein said first modulation technique is a form of QAM.

3. The method of claim 1 wherein said first modulation technique is a form of CAP.

4. The method of claim 3 wherein said first modulation technique is 2ⁿ-CAP, wherein n is an integer.

5. The method of claim 4 wherein n equals 4.

6. The method of claim 1 wherein said multicarrier modulation technique is a form of orthogonal frequency division multiplexing.

7. The method of claim 1 wherein said multicarrier modulation technique is DWMT.

8. The method of claim 7 wherein said first region of the signal frequency spectrum lies below about 6 MHz.

9. The method of claim 7 wherein said second region of the signal frequency spectrum lies between about 6 MHz and 30 MHz.

10. A duplex modulation method which is employed over unshielded twisted pair cable (UTP) in a FTTC system, wherein downstream information is sent over the UTP in one direction to a customer device and upstream information is sent over the UTP in an opposite direction from said customer device, said method comprising: using a first modulation technique to send said downstream information over said UTP and within a first region of the signal frequency spectrum; and using a multicarrier modulation technique to send said upstream information over said UTP and within a second region of the frequency spectrum, wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from said multicarrier modulation technique.

11. An apparatus for transmitting downstream signals over a cable to a remote unit and for receiving upstream signals sent over the cable from the remote unit, said apparatus comprising: a transmitter which employs a first modulation technique generate and send said downstream signals over said cable, wherein said downstream signals are within a first region of the signal frequency spectrum; and a receiver which employs a multicarrier demodulation technique to receive and demodulate said upstream signals, wherein said upstream signals are within a second region of the signal frequency spectrum and wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from said multicarrier modulation technique.

12. The apparatus of claim 11 wherein said first modulation technique is a form of QAM.

13. The apparatus of claim 11 wherein said first modulation technique is a form of CAP.

14. The apparatus of claim 13 wherein said first modulation technique is 2ⁿ-CAP, wherein n is an integer.

15. The apparatus of claim 14 wherein n equals 4.

16. The apparatus of claim 11 wherein said multicarrier modulation technique is a form of orthogonal frequency division multiplexing.

17. The apparatus of claim 11 wherein said multicarrier modulation technique is DWMT.

18. The apparatus of claim 17 wherein said first region of the signal frequency spectrum lies below about 6 MHz.

19. The apparatus of claim 17 wherein said second region of the signal frequency spectrum lies between about 6 MHz and 30 MHz.

20. An apparatus for receiving downstream signals sent over the cable from a remote unit and for transmitting upstream signals over a cable to the remote unit, said apparatus comprising: a receiver which employs a first modulation technique to receive and demodulate said downstream signals, wherein said downstream signals are within a first region of the signal frequency spectrum; and a transmitter which employs a multicarrier modulation technique to generate and send said upstream signals over said cable, wherein said upstream signals are within a second region of the signal frequency spectrum and wherein the first region of the signal frequency spectrum is above the second region of the signal frequency spectrum, and wherein the first modulation technique is different from said multicarrier modulation technique.

21. The apparatus of claim 20 wherein said first modulation technique is a form of QAM.

22. The apparatus of claim 20 wherein said first modulation technique is a form of CAP.

23. The apparatus of claim 20 wherein said first modulation technique is 2ⁿ-CAP, wherein n is an integer.

24. The apparatus of claim 23 wherein n equals 4.

25. The apparatus of claim 22 wherein said multicarrier modulation technique is a form of orthogonal frequency division multiplexing.

26. The apparatus of claim 20 wherein said multicarrier modulation technique is DWMT.

27. The apparatus of claim 22 wherein said first region of the signal frequency spectrum lies below about 6 MHz.

28. The apparatus of claim 27 wherein said second region of the signal frequency spectrum lies between about 6 MHz and 30 MHz.