CA2299559A1 - A power line communications system - Google Patents

Abstract

A power line communications system has a power distribution system including at least one transmission line for distributing power to consumers, a headend communications device connected to the transmission line, and a plurality of distributed customer premise devices connected to the transmission line. The power distribution system serves as a communications bus to permit communication over the power distribution system between any of said customer premise devices and said headend device. The connections are preferably established as virtual connections in an ATM network.

Description

A Power Line Communications System Field of the Invention This invention relates to a communications system, and in particular to a system capable of establishing communication over a power line distribution system, such as a typical electrical distribution network.
Background of the Invention In the classical paradigm, most consumers are served by an electrical distribution system and some form of telecommunications system, such as the public telecommunications network (PSTN). These are entirely separate systems, usually run by different so organizations. The necessary infrastructure is costly to install and involves an inevitable amount of duplication since each system runs separate lines to each consumer.
There have been proposals to transmit communications signals over power lines, but such proposals have to date not met with much success.
Summary of the Invention 15 An object of the invention is to provide an effective means of establishing communications over a power line network.
According to the present invention there is provided a power line communications system comprising a power distribution network including at least one transmission line for distributing power to consumers, a headend communications device connected to said 2o transmission line, and a plurality of distributed customer premise devices connected to said transmission line wherein said power distribution system serves as a communications bus to permit communication over said power distribution system between any of said customer premise devices and said headend device.
The communications are preferably established as virtual connections over a cell-relay 2 s network, such as ATM, asynchronous transfer mode.
In search of an optimum method for organizing data transmission on the powerline, various methods were studied. Due to the topology of the network, a bus system offers optimum conditions to meet all requirements. When defining the transmission method, various secondary conditions had to be taken into consideration, which are essentially determined by the type of signals transmitted.
1. Communication in the power line (PL) network is based on a logical star in whose centre the headend powerline device (HPD), is arranged. The HPD realizes the interface to the conventional telecommunication networks. The interface to the s users of the network or to the telecommunication terminals is produced via the CPD, the consumer powerline.
2. The transmission method on the lowest Plane is realized via a modem. This type of transmission is usually used for data transfer between two devices and not for bus systems. In principle, it can be seen in that various receivers (the CPD) io receive the signals from the HPD and that, vice versa, a series of CPDs must use the same frequency band in a time-division multiplex manner for sending in direction of the HPD. Since every CPD operates on a receiving frequency band that is different from the transmitting frequency band, the access to the transmitting medium must be configured so as to be collision-free by means of is appropriate measures.

3. The PL-bus system should be capable of transmitting voice signals of high quality. That requires maintaining specific minimum transmission rates and maximum delay times.

4. The operational mode of the modem, in particular also its dynamic behaviour, was 2 o practically unknown. However, it had to be taken into consideration from the start that the transmission speeds that can be achieved with the modem would be subject to considerable development. That is, the transmission method had to be, to a great extent, independent from the transmission speed.

5. In view of the fact that the bandwidth currently available in the PL system is 2 s rather small, a method must be selected that enables optimum use of the bandwidth. The optimum must here be looked for especially between the telegram length and redundancy.

6. The problematic or at least unknown interference situation in the network, as well as any possible restrictions when dimensioning the transmitting power, require the 3 o use of repeaters in the bus system.
_Z _ 7. It should be possible to realize the method with low expenditure, whereby the existing design risks should be as minimal as possible.
The aforementioned secondary conditions finally result in the proposal of a transmission based on the cell principle, largely modelled on ATM. An advantage is that a dedicated bandwidth can be allocated to a connection. It is thus ensured that voice signals can be transmitted with the required quality.
Brief Description of the Drawings The invention will now be described in more detail, by way of example only, with reference to the accompanying drawings, in which:-io Figure 1 is a block diagram of a star-based network in accordance with the invention;
Figure 2 shows a hierarchical bus structure;
Figure 3 is a block diagram of a power line device;
Figure 4 shows transmission and reception in a PL network (S - synchronous);
Figure 5 shows device and channel addressing with VPI and VCI;
i5 Figure 6 shows the connection process when the HPD requests a connection;
Figure 7 shows the connection process when the SU requests the connection;
Figure 8 shows the problem of the multiple utilization of line positions;
Figure 9 shows the bandwidths in the various zones dependent on the active SU
channels;
Figure 10 shows illustrates the principle of address conversion during repeating;
2 o Figure I I shows the transmission synchronization and delay times;
Figure 12 shows the zone structure of the network (RPD repeater powerline device;
Figure 13 shows the signalling for start-up of a CPD;
Figure 14 shows the optimal transmitting power and repeater selection; and Figure 15 is a more detailed block diagram of a power line device.
25 Detailed Description ofthe Invention It will be explained in the following how the transmission is organized in the powerline system in a cell-based manner. In principle, the power line (PL) system is distinguished by the following properties:
Referring now to Figure I, the communications system according to the principles of the invention is arranged in a logical star system wherein the headend device (HPD) s communicates over two-way communications channels with the customer premise devices (CPDs). Figure 2 shows how this structure is arranged in a bus-based system, wherein the power line A serves as a communications bus. Media access according to a polling method. The Master is the HPD or, within the a respective subnetwork, the CPD
operating like a repeater.
io Access to the transmission medium refers to use of a specific frequency band or, better, a line position, characterized by modem parameters. In the arrangement shown in Fig. 2, one or more line positions can be used in a zone.
Structure of a Powerline Device As shown in Figure 3, the power line device consists essentially of three functional i5 blocks:
1. The service units ~SU~ A service unit forms the interface to the user of the PL
network. Various services are provided to via a SU. Telephone and data transmission services are provided directly via integrated interfaces (ISDN, POTS). These SU integrated in the PD are called basic service units (BSU).
2o Other services are produced when connections are made at interfaces of the PD
device by third parties, e.g. TK installations.
2. The management unit (MU) This component forms the computer core of a PD, the main task of which is to manage the existing interfaces to the consumer and to the network.
2s 3. PL-Modem (PLMI The layer I corresponding to an OSI reference model is essentially realized by the powerline modem. There are two interfaces to the modem:
The configuration interface is designed like an RS 232C inteface and enables the MU of a PD to transfer corresponding modem parameters with which transmitting and receiving band of the modems can be defined. It is assumed that, although the data are transferred in a delayed manner yet at the same speed to the modem by the MU, as they are delivered to the PL. Moreover, the transmitting power can also be optimized via this interface.
The data are delivered serially to the modem. For example, a Utopia interface can be s employed.
The configured interface can be redefined e.g. as a parallel port or the like.
The powerline network is configured during a relatively complex start-up phase. At the end of the start-up, the available total bandwidth according to the local circumstances and the activated services (SU) is distributed to the PL network. Configuration procedures of io this type must be repeated ~ during activation of SUs, ~ reconnection of CPDs and ~ breakdown of CPDs.
In the following description of the basic functions, we are first proceeding on the is assumption of a static, i.e. completely configured, undisturbed system.
Data Transmission Concept: Basic Principles and Access Methods The HPD transmits a continuous stream of cells. The number of cells per time unit or better the bit frequency is thereby completely independent of the capacity utilization of the PL network. The bit frequency is defined during the start-up phase of the system and 2 o corresponds to the maximum bandwidth available in the system. Each cell, which is typically an ATM cell, consists of the usual 53 bytes which, in turn, differ by a 5-byte header and a payload of 48 bytes. The cells proceeding from the HPD form a time-division multiplex frame for the connected CPD, in which they, for their part, transmit cells to the HPD. These time frames are allocated asynchronously to the CPD
according 2s to their communication requirement. This basic principle is shown in Fig.
4:
The HPD transmits a cell current tuned to the RPD. An explicit tuning of transmitter and receiver is not required in this case or takes place continuously during the transmitting process. Due to the topology of the network and the unknown and undeterministic interference processes on the powerline, however, an additional synchronization process must be included when transmitting from the CPD in direction of the HPD in which only one cell is typically transmitted. This is taken into consideration by the assumption that a synchronous byte (S) is placed in front of each cell. To organize both transmitting directions, from and to the HPD, symmetrically with the same channel capacity, the HPD
s also always inserts a synchronous byte.
These synchronous bytes are not generated by the MU, but by the PLM and also removed by it! Moreover, the PLM requires that the length of the synchronous flag does not exceed a byte duration.
Only connections between the HPD and the CPD are provided. Direct cross-connections Zo between two CPDs cannot be realized. Each cell transmitted by the HPD
addresses a CPD (management information) or a SU within a CPD. It also reacts with a cell on every cell that receives a CPD. Thus, if a CPD receives a cell from the HPD, this results in a transmission authorization of this CPD.
The HPD transmits cyclically to all active SU cells. The resultant cycle can be defined as 15 follows:
A cycle describes the period in which all active SUs receive transmission authorization once. The cycle time is measured in such a way that all SUs can deposit their data in real time.
The number of resultant data is produced, in turn, from the required scanning rate and the 2 o signal release to the connected service. Generally, the following applies to the required bit frequency in a transmission system f~,-&8'~S'f"tf~s-rrij oderausgeklamrtoert fn=&'S-,fi~lf~}
mtt fp- Ai2fl~BCjUBtIL
a - Anxani aer Kan~~le im System s - t3yrte proab~astetem Nu~tzinEormationswcvt t,- AtHast~requenz m - i4larr~gement~ UYefheed gro tSSnat in t3yt2-n - Nut2int01T178dOt1 pro ZeII~ in l3)'~e-wherein fb - bit frequency -z - number of channels in system s - byte per scanned useful information value fs - scanning frequency m - management overhead per channel in bytes n - useful information per cell in bytes For an uncompressed voice transmission, which is at the centre of these considerations, a transmission rate of 8000 frames/second at 8 bits/per channel is usual.
Moreover, it can be assumed that the available bandwidth B corresponds to the maximum possible bit frequency fb. It is known that the management overhead m results from the 5-byte cell Zo header added together with a byte synchronous information. From this point of view, a dimensional equation for the number of channels operable in a cycle follows for the system in question.
e~ m~ rt~c~ ~ ~ ~ 0 gar xB~tsnz ~sr. i~a t~~;~~i l~ ll That is, for 2048 kbit/s, 28,44-28 channels result.
i5 It should be noted that the corresponding bandwidth must be available in each direction.
Moreover, slight shifts can still be made here by integration of D-channel protocols. The following possibilities for better use of available bandwidths are not utilized:
~ voice compression, e.g. to 4 bits per scanning value ~ unsymmetrical structure (no synchronous byte from the HPD direction), in order 2o to thus e.g. realize a simplex management channel ~ sacrificing transmission quality A channel to a SU is allocated asynchronously, i.e. if required and of course depending on availability. If it has been allocated, then the respective SUs are operated synchronously, i.e. also occupy the'powerline when there are no current data in the SU for transmission.
2s (corresponds to the quality feature CBR (constant bit rate). More or less, this condition does not occur in voice transmissions if the cycle time is dimensioned, as just described above, such that the data delivered by a SU can just be transported away.
Addressing Concept The virtual path identifiers (VPI), common in ATM networks, and virtual channel identifiers (VCI) are used for addressing. During the start-up phase, a VPI is allocated to every CPD and just as many VCI (>1) to each SU at a CPD, how channels can be operated by the SUs (So= 2 x VCI). In this way, the device, per VPI, a SU
within a device is completely adressed via the VPI. VPI = 1 is reserved for the HPD.
Moreover, an additional unit is found in the HPD which is only realized in software and which intercepts all calls coming in from the powerline direction. This unit contains VCI=1.
to Fig. 5 illustrates the principle. Basically, the following applies: With cell transmissions from the HPD in direction of the powerline, the receiver is addressed via VPI/VCI. With transmissions in reverse direction, an addressing is not required. All data must be transmitted to the HPD. For this reason, VPI/VCI in this direction contain the sender or the source of a cell.
i5 Beyond interfaces of the HPD, an address translation of VPI/VCI is required into standard telephone numbers. These configuration data are included and managed via the component "Organizational Resource Management" /5/. A table is used in the HPD
that allocates internal call numbers to the corresponding VPI/VCI. The ISDN call number system according to the ITU-T recommendation (E. 164) is used. This number, 2 o consisting of 15 digits, is used as follows for the time being:
Country code max. 3 digits Area code max. 5 digits Subscriber number Part A max. 4 digits External Network 25 Subscriber number Part B max. 3 digits ISDN Sub-address max. 15 digits transparent Therefore, 999 telephone numbers can be provided. If the HPD contacts a CPD in order to exchange management information, then it transmits a cell with VPI =
devices VPI, VCI =0. For general purposes, VPI=0 is also available. Cells provided with this path _g _ identifier are read by all attached PDs. Thus, there is the possibility of transmitting broadcast cells.
Process of Establishing/Disconnecting a Connection (See Figure 6) A connection is established to transmit useful information (SU) or management information between HPDs and CPDs. Both processes do not differ with respect to the establishment and disconnection of a connection. A difference does have to be made between the request for a connection by a CPD or the HPD.
Case A: A call for a subscriber in the powerline network is received at the HPD (See Figure 6) I . The HPD first detects from the call number VPI and VCI of the SU with which a communication is to be established. To the extent that there are still free channels available in the PL system, one of these channels is occupied with the corresponding VPI/VCI and the function message "SETUP"
2. In response, the HPD receives a message "SETUP ACKNOWLEDGE", in the subsequent cycles, the channel continues to be occupied although the connection to the end subscriber has not as yet been acknowledged.
3. While the CPD attempts to establish the connection via the addressed SU, the HPD in the corresponding channel sends the function message "CONNECTION
READY?" and receives a status message "CONNECTION PENDING" or "CONNECTION ESTABLISHED" in response.
4. After the connection has been accepted by the end subscriber, this information can be forwarded to the corresponding SU of the HPD, e.g. the El interface, which can then operate its signalling process with this information. At the same time as this "established", a channel for the data transmission is available in the PL
system.

5. Finally, the SU on the CPD terminates the conversation or the connection. The corresponding CPD then sends "DISCONNECT" on the powerline. If the message comes from the HIDD, then it sends the message "DISCONNECT", receives "DISCONNECT ACKNOWLEDGE" in response and thus terminates the connection, the channel is released. If the connection at the SU of the CPD is terminated, the signalling of "DISCONNECT" is sufficient.
Case B: A call for a subscriber outside of the powerline network is received at a CPD
to (see Figure 7) I . In view of the fact that transmitting activities are not allowed for the CPD on its own initiative, it must first wait until the HPD requests that it transmit status information. In all free channels that are not occupied by active connections, the i5 HIDD requests all connected PDs to send their status to the HIDD. If no channel is free, then, at any rate, no new connection can be established. A connection request is then not accepted.
2. The message "SETUP ACKNOWLEDGE" is sent in response to the function 2o message "SETUP", while the signalling process takes place in the overriding network. The CPD then asks with "CONNECTION READY?" for an established connection, in response to which "CONNECTION PENDING" is sent as long as there is no connection. This cycle is repeated until it is interrupted by a "CONNECTION ESTABLISHED" or a time-out.
3. After the connection has been established, the CPD begins with the transmission of data cells.
- to -5. The rules set out under A apply for termination of the connection:
Depending on which device the connection is terminated, a "DISCONNECT" is sufficient or it must still be answered with "DISCONNECT ACKNOWLEDGE".
The function messages are transmitted as a 16 bit code word. The other 46 bytes of a cell s are available for further information. Depending on the type of message, telephone numbers, type of connection or the like are also transmitted /7/. The HPD
monitors exceeding time-out parameters in each phase of the connection.
Expansion of Data Transmission in n-Zones As could already be seen in Fig. 2, the range of a HPD is increased at relatively low so transmitting power by the use of repeaters. Repeaters are normal CPDs that are defined by the HPD during the start-up process. In principle, a CPD has 2 modems so that every device can be used as a repeater.
The first modem is used for data exchange with the HPD (or another repeater).
The second modem 2 is used for the repeating. Ever repeater (RPD repeating powerline i5 device) marks the transition into a new zone, with which no message concerning the topological situation is associated. Only those signals that have their recipients in subordinated zones are regenerated. That is why the expression "repeater"
should not be understood to mean the same as in network technology. In fact, this is a bridge. It is fundamentally assumed that the modem parameters/line positions preset by the HPD or a 2 o RPD may not be used again at any point in the network. Although such a multiple use of line positions is possible in principle, it does, however, require a considerable additional set-up expenditure. The basic problem is shown in Fig. 8. The linear damping of the signal accepted there results therein that the line position is recognized as free at CPD 3.
However, if CPD 3 begins to transmit on this line position, then this signal acts as 2s interference at CPD 2. Thus, the use of a multiple utilization strategy requires extensive tuning procedures between all PDs (CPDs and HPDs). Whether a multiple utilization of bandwidth is possible, can in any case scarcely be calculated in advance and depends greatly on the actual network conditions. For this reason, the following 2 developmental steps are realized first.

The same number of channels are offered in all zones This method is very simple to realize. The number of channels provided by the physical layer is distributed uniformly over the zones. The following applies:
.~xk~srar It..
$.t: L~' k wherein Bsys - max. bandwidth in the system BX - required gross bandwidth of a channel KSys - max. possible number of channels in the system so fb - bit frequency and for the channels realizable in a zone tic e;~ . ~r~x. aarr~~; elite iris .rte. rem g,, - ~rfor~rficl~ 8~ruttoA~rn~bneite eines K~rn~t~
yp, -max. rscb~l~ft~ Afli9FYl tl~rVfArlAl? im SystPm fb ~ Ciirfre~uenz erasl tiir 47~i: Ira fir ~r~~t~ ~e~lisirtria~trrr~ ~4;rrt~ie -'~~ t~sit2~A11xarh14~l3On~rtitnS~tciY1 I~,..'a~-with Z number of zones in the system In all determinations, the effectively useful bandwidth is halved with each repeater when i5 using this method. A differentiated observation based on an actual load to be expected does not take place. However, the method is simple to control and absolutely fair with respect to the control probability of a waiting channel.
Graduall~arovided bandwidth In all zones n, only those channels are provided that are not already reserved for 2 o transmission according to a method to be determined in the zones n - 1 ...
1. At the same time, the channels provided in zone n are not occupied in zones n - 1 ... 1.
Accordingly, [31 is the number of channels in the system having n zones:

t a '.y's ':
~.~ x ~ ~ l ,~t~5 ~J4Y~~dl ~1"'t~s;9t1 f!~1-.
with n = number of zones The overall channels available per zone are divided in the ratio of active terminals/SU
channels. For this purpose, the formula is used.
S~ active channels in the SU of zone 1 s, rel. portion of zone 1 in the active channels in system S~ active channels in zone I
1 o A valence of the individual zones also results from the ratio of the SU
channels available in a zone to the sum of all available channels. Accordingly, it should also be true that there should be proportionality between the number of SU channels and the number of useful PL channels between the zones.
~v_'~a~
K,, S~
15 If one changes this equation to K q and uses I = 1 and inserts sys for K in the summation formula, then the following applies fC~,,~~~K ~+2 ~~ ~ K, +3 ~~ 1~~ ... n ~° I~ ~
,~~ ,$f ~t ,K~ 1-+25~~+3 ~3 ..rr~°' ~ J:~[1-+'~iS, r.,~ ~, For the dimensioning of K, this results in i l _h S, ~ i,Sr) Depending on how the number of active service channels are distributed in the system, a s different picture results with respect to the distribution of channels.
Example:
2 repeaters must be used in a network. The available bandwidth is 5 Mbit/s per direction.
Zo In zone l, there are 30 active SU channels, in zone 2 - 40 and in zone 3 -50.
The channel number of the system is as per [3]
3~btrPs JC~x ~ Is °~tlCa'Xtlg'ICA' t~Ql~~t~
~~allt~lE
-~ I ' dri-z~t~~~~r3~t:~~yl ~,~~~~, channels channels ~~~,~.~,3t~~!?,i~.~ia Thus, a distribution of the 64 available channels over the zones results as per Fig. 9.
Whether a SU is occupied more or less is not considered in this procedural manner.
The frequency with which a SU cannot occupy a channel is used as a standard for the 5 adaptation of this method.
This information is collected and used for evaluation and possible change of the distribution code.
If Vi is the occupation frequency in zone i, then the system having n zones is deemed optimally dimensioned when ~ tf a ~si TI'~m applies. If this is not the case, then zones having higher occupation frequency must be furnished with more channels by evaluating the situation on the basis of occupation frequency:
~,.,-~; ~f,.
_~
EXam~le The channel numbers K1 = 7, K2 = 10 and K3 = 12 were ascertained for a 3-zone network. After several days of operation, V 1 = 15, V2 = 45 and V3 = 18. V2 is obviously higher. The evaluation shows:
- IS -K~.~73~~ +~6tE&=4.U~~
~=-~03t~ +45+~$~~7a1'7 ~~~~~2315 -t'4S+tR~8,3~~$
~ ~,irtt ~~utliclh, dues ~tt;
'f fi ~.
i61+K:+tC3 = K3'+K~'+K3'od~all~erm~ln ~i6'.f~~.~;
' t.:l J;".~
It becomes clear that applies or, generally, s Any rounding errors can be avoided with this formula. The new parameters are set by the system. The counters for the occupation frequency are reset and subsequently not observed any further. As must happen often, conclusions must be inferred from the total number of calls conducted in an observation period.
Address Conversion in Repeaters io If a CPD is defined as a repeater, then all VPI/VC1 combinations must be transferred to this device which must again be transmitted via modem 2 with modulation parameters also to be advised or received from there and to be transmitted via modem 1.
Every CPD
has 2 tables per modem in which the relation between the VPINC1 and the ports are defined via which one receives or transmits. A CPD makes 3 ports available:
15 Port 0 = modem 1 Port 1 = modem 2 Port 2 = local This means that all connections existing within a CPD are made via port 2.
Fig. 10 illustrates these correlations.
2 o Delay Times and Synchronization Transmitting from HIPID in direction of the PL network takes place, as described, continuously. The receiving direction of all CPDs should also be capable of detecting this signal continuously and to compensate small fluctuations in frequency or phases.
Above all, the transition to transmitting in a CPD is important for the synchronization.
s The synchronization of the transmitting start is made possible by the receipt of a cell intended for the MU or a SU. The following delay times occur in the transmitting and receiving channel:
Delay times Duration at f'=
2Mbit/s Demodulation times: ~ 15.25 ~
The processes occurring in the PLM are not known in detail. However, it can be assumed that about 1 byte duration, i.e. 8/f~'; passes between the arrival of a bit and the output of this bit to the bit rate adaptation of the MU. In addition, the synchronous byte must be cut off. Thus, a demodulation delay of tli=161fh can be expected.

Delay due to bit rate adaptation 214.12 ps The serial data flow is read in with the bit frequency Ih and first pushed through a 5-byte long slide register. This slide register is used for header synchronization since the 5th byte is compared with the remainder of module 2 division of the first 4 bytes of the header. If this so-called HEC agrees with the rest, then a cell start is detected. After complete reception of the cell, its bytes are transferred at a speed of 25 Mbytes in a byte serial manner via the Utopia interface to the communications processor. This transfer only takes place if a VPI that is identical to the device VPI or the broadcast VP1 (VPI=0) is detected in this subassembly. The delay time is then tribe=(m+n-1)-8-ll.fh+(m+n-1).llfut'pi, or in summary trine= (m+n-1)- (8/fb + 1/f~topi~
The cell length (m+n) should be reduced by 1 because it is assumed that the synchronous byte in the modem will be removed.
Reaction Time of the MU 8 ps After the cell has been read comPLetely via the Utopia inteface, the communicatons processor (CP) releases an interrupt. The ISR analyzes the header, releases a decision process for the sequence with which the individual channels may send their cells and switches the task to the local further processing of the receiving data to ready. The delay time results solely from the time between interrupt and the return of the ISR.
The ISR is extremely small and is programmed in such a way that it cannot be interrupted. Moreover: The cells must always be ready in the CP for the actual connections and also for the "STATUS" queries.

Bit Rate Adaptation in Transmitting Direction 2.59 ps The data are again incorporated in a byte serial manner via the Utopia interface, controlled by the CP. A speed of 25 Mbytes is thereby realized. The cell is converted into a bit-serial current, the HEC is determined per modulo 2 division on the fly.
The first bit at the output of the subassembly appears with the serialization, insofar as the PLM

indicates ready to accept, immediately after a further pulse in bit frequency. This results in a delay of fdbe=(rn+n- l~lf"topia+1/fb Modulation Time ~p.s The synchronous byte can be transmitted immediately upon arrival of the first bit from the bit rate adaptation component.
For this reason, the delay until occupation of the medium is slight.

The delay times thus essentially depend on the bit frequency, the speed of the Utopia interface, the processor frequency (ISR) and the parameters defined by the modem.
s Unfortunately, it cannot be assumed that all these processes run in a determined manner.
It was therefore specified that the transmitting process already be started from the subassembly for the bit rate adaptation. For this purpose, one proceeds according to the following principle:
io 1. The header of a cell is decoded within the bit rate adaptation. When recognized (HEC = CRC remainder), the transmitting process is released one system pulse later.
2. 2 transmitting buffers per modem are realized within this subassembly.
3. Buffer 1 is fed by the Utopia interface. The MU again decides permanently about the next transmitting cell and transfers data to this buffer when the bit rate adaptation per signal indicates ready to accept.
4. Buffer 2 always contains a status cell. This status cell is again described cyclially (about every 100 ms) via the processor orbus (60x).
5. If the bit rate adaptation recognizes a status query (VPI=device VPI, VCI=0), then the transmitting process starts from buffer 2. This is also true when there are no valid data in buffer 1 and a transmitting authorization is granted.
For this purpose, it is not necessary to respond to a cell that was sent from the HPD to a specific SU with a cell from the same SU. Only in the sum, all addressed SU
channels Zo and the MU must react in a real-time capable manner, i.e. deposit cells in direction HPD.
If transmission is started by the FPGA of the bit rate adaptation, then this process can be considered to be determined. By interference in the modem, the decision about the media access can perhaps be relocated to layer 1. As a result, the defining methods would be simplified or become unnecessary. Due to new recognitions with respect to functioning of the modem, especially the operation of the bit rate adaptation is greatly affected. As a result, shifts of delays from one subassembly to another become possible.
Fig. 11 clearly illustrates the resultant consequences:
1. Cell A arrives at the PD, is delivered to the bit rate adaptation which decodes the ao header and detects a relevant VPI.
2. As per VPI/VC1, data are delivered to the modem from buffer 1 or 2. The transmitting process is released with the first flank of the data delivered to the modem. The modem sends the synchronous byte or the like.
3. The cell is delivered to the CP, the transmitting scheduler, the MU, which decides from which buffer a transmission is to take place next, is activated. If present, 48 bytes of data and a 4 byte header consisting of VPI and VCI per Utopia interface are delivered to the bit rate adaptation. If there are no data present, nothing happens.
The following time elapses between the arrival of a cell at the CPD and the start of the transmitter in direction HPD:
td-transmit-tdd'~-tth'~l~fb+tdn With tdH delay until comPLete decoding of the header fb so This, in turn, results from 40 bit times corresponding to the header length of the cell. At a bit frequency of 2 Mbit/s, therefore, a delay time of 20.008 f~ s results.
Attention: The synchronous byte prevents, on receipt, conflicts between the reading in of the data in bit frequency and delivery of the data per Utopia. No data can be accepted at the modem interface while the Utopia interface is active. This is caused by the different pulsing at i5 both interfaces of the bit rate adaptation. This must be taken into consideration when shortening synchronization times.
The complex, electrically inhomogeneous network requires an intelligent start-up strategy that must be suitable for forwarding PDs for utilization without great manual interruptions and, above all, without knowledge of the physical structure of the network.
Any number 20 of CPDs desired and a HPD exist in every newly constructed PL network.
Since the topography and the interference situation of every network is different, all requirements must be met during start-up to maintain the communication between HPD and the CPDs under all operating conditions. At the end of the network initializations, the transition is at normal operation.
25 Course ofthe Start-up Process The purpose of the initialization process is to ascertain optimum transmission paths within the low-voltage network that are defined by corresponding modem parameters.
An optimum transmission path distinguishes itself thereby that a CPD with the necessary bandwidth is attained at minimal transmitting power. The necessary bandwidth corresponds to a standard bit rate that results from the number of channels to be operated simultaneously in a network zone multiplied by 64 kbit/s. If one looks at Fig.
12, it becomes clear that the maximum possible bandwidth must be available in zone 1.
Individual CPDs can be used as repeaters, as already noted. The number of repeaters s should be as low as possible in the interest of minimizing additional delay times.
The initializing process starts at the HPD. Every CPD receives and sends with preset modem parameters that are distinguished by a specific line position, a very low bit rate (8 kbibs) and a maximum transmitting power. It may be necessary to correct in such a way that every connected PD is securely obtained with these parameters. As a prerequisite, so the HPD has the list of connected CPDs. The HPD starts after it is switched on in the unconfigured state by sending a continuous ATM cell current in which it is addressed with a status cell (WAKE UP) for and after every CPD. The cell current is filled with empty cells to give the CPD time to react. For this purpose, every status command is repeated exactly after 20 empty cells. If the CPD also does not respond to the second i5 status command, then the HPD addresses the next CPD and will again try later to address the device not obtained via a repeater to be defined. The "WAKE UP" cell contains a PDU, in addition to the function command, with the VPI which is allocated to the CPD.
Since the CPD cannot as yet be addressed directly, VPI=0 and VCI=serial number of the CPD. Cells having such a header are only received during start-up. The CPD
responds to 2o the WAKE UP with its configuration data set and an identification number for the quality of the reception ("WAKE UP ACKNOWLEDGE"). From the configuration parameters, the HPD detects the maximum bandwidth to be provided for this CPD and reserves a corresponding number of VCIs. The PD is now requested to switch to a specific line position. For this purpose, the HPD sends a "TRANSMISSION CHECK", whereby 2s VPI= "device VPI" and "VCI=0" is now already used. The CPD is thus advised on which line position (modem parameter for modulation, transmitting power, bit rate) it should expect a test cell current. In addition, the configuration data for the SU (SU
code, VCI, active/inactive) are found in the PDU.
The CPD must now define modem 2 on the desired line position, modem channel 1 3o remains in the line position for the initialization channel.
The CPD quits this command (TRANSMISSION CHECK ACKNOWLEDGE) via the initialization channel.
The HPD begins with the transmission of the test cell current whose payload consists of significant bit combinations. The CPD permanently evaluates the bit error rate and in turn sends a test cell current to the HPD, where the bit error rate is also evaluated. If a s transmission takes place at all, then 10,000 test cells (530,000 bytes) are sent. The HPD
then queries with TRANSMISSION RESULT, the result of the test at the CPD end.
The CPD responds TRANSMISSION QUALITY, to which the HPD redefines, if necessary, the transmitting power of the HPD and CPDs within the scope of possibilities (TRANSMISSION CHECK). The procedure described is repeated until an optimum io transmitting power is found in which the necessary bit error rate is just obtained. Ifthere is no transmission, then a new line position of its predefined transmission parameters (8 kbit/s) is defined for modem 2 via modem 1 Lind, with which the attempts to establish a transmission having the required bandwidth is repeated. Fig. 13 illustrates the described signal play. The values established after successful conclusion of the tests for line i5 position, transmitting power and bit frequency are stored in a configuration table within the HPD. The bit frequency is then increased and the same process completely repeated, until the maximum number of channels provided is obtained (at 2,048 Mbit/s as channels). Finally, the HPD defines the transmitting and receiving parameters of the CPD for the operating status.
A Table is produced in the HPD for every CPD having the following form:
Modulation Transmitting power Bit Rate The registered values result from the allocation of firmly defined parameter sets (modulation) or from the standardization of the corresponding absolute values 2 s (transmitting power 0... 1000, bit rate corresponding to number of channels). During these comprehensive tests per CPD, tests with other modem parameters and the same bit rate are also repeated with positive test results. In this way, the HPD has comprehensive knowledge with respect to the physical conditions on the powerline, insofar as these have an effect on the transmission quality. If interferences occur later in a connection during the running operation, then these data can be used quickly and without new tests for locating replacement transmission paths /I/.
CPDs that cannot be obtained after termination of the described tests or not with sufficiently high bit rates, must be connected via repeaters.
Selecting a Repeater so After all parameters for the CPD have been detected within zone l, the function of the HPD is to search for all stations which were not found with the maximum signal of the HPD or in which the error rate was not slight enough. To this end, it successively uses the stations thus far obtained as repeaters.
The task of the selected repeater is to regenerate the signal received in such a way that the is devices not yet obtained now receive an optimum signal by division of the path length, whereby all participants attempt to minimize the transmitting power. The aforementioned algorithm is again used, whereby the RPD is charged as a substitute With the transmitting and receiving of test cells through the HPD. The slow initializing and management channel remains under complete control of the HPD. A modem of the RPD also remains a o connected with the HPD via this channel just as with the OPD which is now addressed via the RPD.
A repeater is defined by "REPEATER SETUP" under "VPI=Devices" VPI and "VCI=0".
With this command, it receives a list of serial numbers of the OPD that are to communicate via the repeater and a VCI for the repeater. As usual, the OPD
responds 2s with "SETUP ACKNOWLEDGE" and immediately begins with the known procedure for establishing a connection with the OPD (serial number list). The contact to the HPD is ensured by the slow initializing channel via which the HPD permanently requests status information from the RPD via the HPD.
A CPD is selected as potential repeater based on the parameters determined in zone 1.
3 o For this purpose, the average optimum power is detected in the already known network and compared with the transmitting power for obtaining the other OPD in zone 0. The average optimum power is then calculated as follows:
f~r~t~r s wherein m = number of CPDs obtained. PoPt(n) = optimal transmitting power CPD
>HPD.
The deviation of the individual OPDs from the average results therefrom with:
~~~nycLR~-~~r~u'..~~etral A sorted list of these deviations is produced and the OPD first selected as repeater which to has the lowest power deviation from the average optimum transmitting power.
With this, the OPD is first used as repeater which presumably lies in the central field of the obtainable network. It can thus be avoided that the border devices that are either very close or very far removed from the HPD are accidently used as repeater. As a result, in particular, a minimization of the transmitting power is obtained, at the same, options is result for possible error strategies.
Activation of a new CPD
The activation of an individual CPD takes place, in principle, in the same way. As a prerequisite, the serial number of the new device is present at the HPD so that, via the initializing and management channel, "WAKE UP" is cyclically sent. The new device is 2o connected either via a repeater or directly into the PL network.
Power line Device A more detailed diagram of a power line device is shown in Figure 15. This comprises a microprocessor connected to a power line interface through a Field Programmable Gate Arrar (FPGA) to an Rs232 and 10/100 Base T Ethernet interface.
2 s The microprocessor MPC is also connected to a RAM storing data and instructions, and flash memory Flash, and a PCI bridge.
Ports SMC l, 2, 3 are connected to Phy interfaces, such as Utopia interfaces.
The PLM is connected with the basic assembly via a plug connector and mounted on it as a piggyback assembly. The interface is yet to be defined structurely and with respect to s the plug engagement. The PLM can be provided with SV via the basic assembly.
The required power is briefly specified.
RS 232 (1) Monitor interface for inspecting logging RS 232 (2) Configuration interface for the PLM 1 and 2 Utopia I ( 1 ) Data interface for the PLM I
s o Utopia I (2) Data interface for the PLM 2 SAR (I) Alternative serial interface for PLM I
SAR (2) Alternative serial interface for PLM 2 Ethernet Interface for future LAN services So( I ) Interface for local terminals i5 So(2) Interface for local terminals El Interface for local terminals The described invention thus provides a device that can provide utilities with the ability to offer various service interfaces in the weak-current range. Standard telephone interfaces are available as service interfaces. The device makes it possible to conduct telephone 2o conversations of standard quality via power-line connections. In addition to the aforementioned functions that are made available to the user of the system, the device provides the transition from the power-line to existing networks. Dependent on the arrangement of the device at a user's end or at the transition point to other networks, a distinction is made between the consumer's power-line device (PD for short) and the head 2 s power-line device (HPD).
A standard ISDN telephone can be connected to a PD. If a user dials a standard telephone number, then the PD establishes a connection to the desired subscriber, signals the call and busy tone to the user in the usual manner and realizes the voice transmission between the two subscribers when the connection is accepted. Instead of the connection of a standard telephone, the user can also connect a conventional extension to the PD. If data devices are connected to the telephone interface, then these devices can transmit the data transparently with aid of the PD. A certain service quality (QoS), in particular bandwidth, can also be made available to the user, especially for the transparent use of the PDs for data services. The PD or all participating PDs prepare all information that is required for calculating the cost the telephone service.
In addition to voice transmission, the PD realizes the transmission of data and thus offers further operational possibilities, such as the provision of an Ethernet interface. Data and Zo voice transmissions are differentiated at the point of the transition into other networks in such a way that they can each be delivered to other networks. It is thereby irrelevant whether the data are fed into the system via a telecommunication interface (ISDN) or a data interface.
The product is used in the low-voltage range of national grids. Use is largely is maintenance-free. Operation by the user is not necessary. A person/machine interface is only required to a limited degree and exclusively for trained technical personnel. Third parties should not have unauthorized access to a PD or to the entire power-line network.
The PD is typically used in those areas in which the residence connections of the national grid are found. Temperature range and other parameters are to be designed accordingly.
2 o At a Consumer PD, the device typically (a) Provides 2 European ISDN connections for the connection of telecommunication transmitting devices, especially telephone sets and telephone installations.
Forwarding a connection request in the power-line network to the HPD. Establishment, operation and disconnection of up to 4 parallel telephone calls, to the extent that the bandwidth of the 2 s overall system permits this.
(b) Provides a S2M/ El connection for connection with standard telephone installations.
Realizing telephone calls via the telephone installation. Establishment, operation and disconnection of up to 30 parallel telephone calls, insofar as the bandwidth of the overall system permits this and that channels are not already occupied by the existing base 3o connections corresponding to (a) (max. 30 parallel calls per PD).

(c) Provides fee information for conventional telephone systems as well as further features of the ISDN protocol. Paragraphs (a), (b) and (c) should be understood as sections of the device development. Its chronological arrangement in the overall development can be found in the schedule.
s At the headend, the invention can provide (a) 2 European ISDN connections for the transition into external telecommunication networks. Establishment, operation and disconnection of up to 4 parallel telephone calls, insofar as the bandwidth of the overall system permits this. Recording the necessary information for calculating charges.
so (b) a S2M/ EI connection for connection with external telecommunication networks, insofar as the bandwidth of the overall system permits this. Recording the necessary information for calculating charges.
A PD is started by connecting at the interface to the network and to the services as well as switching on the network part. The installer can then test basic device functions via the i5 RS 232 service interface. Moreover, he must determine whether the PD is used as a head station. If necessary, corresponding software modules are added to the equipment software via a service interface. Optionally, it is checked whether the software of the HPD and standard PD have room in parallel in a device. In this case, every PD
should operate as a consumer PD if it is not defined as a HPD via the service interface.
2 o The network is started without knowledge of the network topology, the interference situation and other physical variables. However, it is assumed that all PD and the HPD
are installed properly. The goal of this initializing process is to ascertain the optimum transmission paths within the low-voltage network that are defined by appropriate modem parameters. An optimum transmission path is distinguished thereby that a PD
having the 2s necessary bandwidth is obtained at minimum transmitting power. The necessary bandwidth corresponds to a standard bit rate which results from the number of channels to be operated simultaneously multiplied by 64 kbit/s. The fact that it is in the start-up mode can be ascertained by the LED on a PD. If use of the network is not possible to the desired extent (number of PDs or users), then this information must be provided for 3 o evaluation. Every PD has two transmitting and receiving units independent of one another with 2 power-line modems. The function of a PD as a HPD or consumer PD
is defined by the firmware.
Behaviour of the HPD: The HPD realizes various processes that are suitable for defining transmission paths and which are to be redefined when there are dynamic changes in the s network "on the fly". It has master functionality for all communicaton processes.
Behaviour of the standard PD: Under the guidance of the HPD, the PD can be adjusted with respect to modem parameters. Every PD can, in addition, assume the function of a repeater.
A connected telecommunication transmitting device reacts as usual in today's io telecommunication networks: After lifting the receiver, a dial tone is emitted by the PD.
- The user dials a telephone number. ABer the dialing process is completed, the system attempts to establish a connection for the user. This first takes place via the internal, i.e.
power-line, network and then via HPD in existing standard telecommunication networks.
If the connection is successfully established, then the ring signal is signalled as usual to 15 the user. If no connection can be provided within or outside of the power-line network, then the user receives a busy signal. A telephone connection is realized with good quality comparable to an ISDN connection. The power-line is subject to strong dynamic interferences as known from conventional telecommunication networks or other networks. For this reason, the components (HPD and consumer PD) participating in a 2o connection must also react dynamically to interferences and also switch to alternative transmission paths without considerable interferences during established connections.
After the connection is terminated by the user, all transmission channels are disconnected securely.
The service made use of by the user is recorded accurately. These call data are 2s transmitted via the telecommunication network to a external data bank server.
It can be seen by the LED on a PD that it is in the normal mode.
The PLM is provided with - a bit serial data flow, devided into transmitting and receiving direction, or - an interface corresponding to Utopia 1 as data interface.
The modem parameters, bit rate and transmitting power are defined via a RS232 or another interface that should be specified briefly.

Claims (9)

1. A power line communications system comprising a power distribution network including at least one transmission line for distributing power to consumers, a headend communications device connected to said transmission line, and a plurality of distributed customer premise devices connected to said transmission line wherein said power distribution system serves as a communications bus to permit communication over said power distribution system between any of said customer premise devices and said headend device.

2 A power line communications system as claimed in claim 1, wherein said customer premise devices are connected to said bus in a hierarchical configuration.

3. A power line communications system as claimed in claim 1, wherein groups of said customer premise devices are connected to a repeater device connected to said bus.

4 A power line communications system as claimed in any one of claims 1 to 3, further comprising means for establishing virtual connections between said headend device and any of said customer devices to permit communication therebetween.

5. A power line communications system as claimed in claim 4, wherein said virtual connections are established by means of data carrying cells transferred over said communications bus.

6. A power line communications system as claimed in claim 6, wherein said cells are ATM cells.

7. A method of establishing communications over a power line distribution system comprising connecting a headend communications device connected to a transmission line for distributing power to consumers, connecting customer premise devices to said transmission line at distributed locations wherein said power distribution system serves as a communications bus, and establishing communications between any of said customer premise devices and said headend device over said communications bus.

8. A method as claimed in claim 7, wherein said communications are established as virtual connections.

9. A method as claimed in claim 8, wherein said virtual connections are ATM

connections.